Baixe o app para aproveitar ainda mais

Prévia do material em texto

Ayman F. Habib MMT 2013 – Summer School (Taiwan) QA/QC of Photogrammetric and LiDAR Mapping Chapter 4 Photogrammetric & LiDAR Mapping Ayman F. Habib 2 Overview • Motivation • Quality Assurance (QA) and Quality Control (QC) – Introduction – Prerequisites • Photogrammetric Principles • QA/QC of Photogrammetric Mapping • LiDAR Principles • QA/QC of LiDAR Mapping • Concluding Remarks Photogrammetric & LiDAR Mapping Ayman F. Habib 3 Motivation • There has been a significant advancement in the remote sensing and mapping technology. – Digital cameras provide an alternative to conventional large format analogue cameras for rapid data collection. – Direct geo-referencing is providing the means for an almost control-free mapping environment. – LiDAR provides a dense point cloud representing the object space surface, and thus offers a fast and accurate way of obtaining a Digital Surface Model (DSM). • Effective utilization of these advances mandates the development of reliable, practical, and standardized procedures for the Quality Assurance (QA) and Quality Control (QC) of the mapping process. Photogrammetric & LiDAR Mapping Ayman F. Habib 4 • Quality Assurance and Quality Control are important concepts, yet we might have only a vague understanding of the meanings and the differences between these terms. • Quality Assurance (QA) refers to the process used to create the deliverables. – QA includes process checklists for project audits. – If a process is audited, an auditor might not be able to tell if the specific deliverable is acceptable (QC). – However, the auditor should be able to tell if the deliverable seems acceptable based on the process used to create it (QA). • Quality Control (QC) refers to activities associated with the evaluation of project deliverables. – QC is used to verify that the deliverables are of acceptable quality and that they are complete and correct. Quality Assurance & Quality Control Photogrammetric & LiDAR Mapping Ayman F. Habib 5 5 • Quality Assurance (pre-mission): – Management activities to ensure that a process, item, or service is of the quality needed by the user – It deals with creating management controls that cover planning, implementation, and review of data collection activities. – Key activity in QA is the calibration procedure. • Quality Control (post-mission): – Provide routines and consistent checks to ensure data integrity, correctness, and completeness – Check whether the desired quality has been achieved Quality Assurance & Quality Control Photogrammetric & LiDAR Mapping Ayman F. Habib 6 • To develop effective QA/QC procedures, we need to understand the mechanism of the mapping process including: – Data acquisition systems, – Error sources (random and systematic), – How to mitigate the impact of these error sources, – Nature of available data, – Data processing algorithms, and – Nature of delivered product. Quality Assurance & Quality Control Ayman F. Habib MMT 2013 – Summer School (Taiwan) Photogrammetric Mapping 7 http://earth.google.com/ Photogrammetric & LiDAR Mapping Ayman F. Habib 8 Frame Cameras Applanix DSS 439 SONY 717 RC30 Kodak 14n Canon EOS 1D DMC IKONOS/GeoEye Line Cameras ADS 80 Medium-Format Digital Cameras Photogrammetric Data Acquisition Photogrammetric & LiDAR Mapping Ayman F. Habib 9 Multi-Camera Systems A rigid-relationship among the cameras Photogrammetric Data Acquisition Airborne Mobile Mapping System Photogrammetric & LiDAR Mapping Ayman F. Habib 10 Multi-Camera Systems A rigid-relationship among the cameras Photogrammetric Data Acquisition Airborne Mobile Mapping System Photogrammetric & LiDAR Mapping Ayman F. Habib 11 Photogrammetric Data Acquisition Multi-Camera Systems A rigid-relationship among the cameras Terrestrial Mobile Mapping System Photogrammetric & LiDAR Mapping Ayman F. Habib 12 Photogrammetric Data Acquisition Multi-Camera Systems A rigid-relationship among the cameras Portable Panoramic Image Mapping System Photogrammetric & LiDAR Mapping Ayman F. Habib 13 Object Point (A) Conjugate Points • The interior orientation parameters of the involved cameras have to be known. • The position and the orientation of the camera stations have to be known. a a´ Photogrammetric Point Positioning Camera Calibration (IOP) Geo-referencing (EOP) Photogrammetric & LiDAR Mapping Ayman F. Habib Objective: Determine the Interior Orientation Parameters (IOP) of the camera The defined bundle by the IOP should be as similar as possible to the incident bundle onto the camera at the moment of exposure. Principal Point Coordinates Principal Distance Distortion Parameters Image Point Perspective Centre Principal Distance c xp yp Principal Point Coordinates Distortions Photogrammetric Camera Calibration 14 Photogrammetric & LiDAR Mapping Ayman F. Habib 15 • Exterior Orientation Parameters (EOP) define the position and the attitude of the image bundle relative to the ground coordinate system. – The position of the bundle is defined by (Xo, Yo, Zo). – The attitude of the bundle (image/camera coordinate system) relative to the ground/mapping coordinate system is defined by the rotation angles (w, f, k). • EOP can be either: – Indirectly estimated using Ground Control Points (GCP), or – Directly measured using GPS/INS units onboard the imaging platform. Photogrammetric Geo-Referencing Photogrammetric & LiDAR Mapping Ayman F. Habib 16 Ground Control Points Tie Points Photogrammetric Geo-Referencing Indirect Geo-Referencing Photogrammetric & LiDAR Mapping Ayman F. Habib 17 Photogrammetric Geo-Referencing Indirect Geo-Referencing Signalized Targets Photogrammetric & LiDAR Mapping Ayman F. Habib 18 Photogrammetric Geo-Referencing Indirect Geo-Referencing Natural Targets Photogrammetric & LiDAR Mapping Ayman F. Habib 19 Direct Geo-Referencing Photogrammetric Geo-Referencing Photogrammetric & LiDAR Mapping Ayman F. Habib 20 c distyy distxx ypi xpi GP XG ZG YG Object point (I) 1. Mapping Reference Frame 2. Image Coordinate System m c r x y z m I r c ii rS c i r ),,(R m c kw Image point (i) c i m ci m c m I rRSrr ),,( kfw Photogrammetric Positioning Indirect Geo-Referencing Photogrammetric & LiDAR Mapping Ayman F. Habib 21 GP xb yb zb XG ZG YG Object point (I) 2. IMU body frame 1. Mapping Reference Frame 3. Image Coordinate System x y z )t(r m b )t(R m b b c r c distyy distxx ypi xpi c i r c ii rS b c R m I r Image point (i) c i b c m bi b c m b m b m I rRtRSrtRtrr )()()( Photogrammetric Positioning Direct Geo-Referencing – ISO Photogrammetric & LiDAR Mapping Ayman F. Habib • Photogrammetric quality assurance includes: – Percentage of overlap – Percentage of side lap – Flying height – Base-height ratio – Number/distribution of tie points – Number/distribution of GCP – Scanning resolution (analog images) – Geo-referencing procedure – Camera calibration – System calibration Photogrammetric Quality Assurance 22 Photogrammetric & LiDAR Mapping Ayman F. Habib 23 • Camera Calibration. – Laboratory calibration – Indoor calibration – In-situ calibration • Total System Calibration – Camera calibration, and – Mounting parameters: spatial and rotational offsets between various system components • The DPRG has been investigating: – Automated procedures for indoor camera calibration – Flight configuration with minimum control requirements for camera and mounting parameter estimation Photogrammetric Quality Assurance Analytical calibration Photogrammetric & LiDAR MappingAyman F. Habib 24 Traditional Versus Proposed Indoor Test Field Traditional test field Test field proposed by the DPRG Photogrammetric & LiDAR Mapping Ayman F. Habib 25 Linear Features & Camera Calibration • Deviations from straightness in the image space are attributed to various distortions. • Incorporating linear features in the adjustment procedure would allow for the calibration of the implemented camera. Photogrammetric & LiDAR Mapping Ayman F. Habib 26 After Calibration • When compared to traditional point-based calibration approaches, linear features proved to be more effective in distortion removal from imagery captured by cameras with significant lens distortion. Photogrammetric & LiDAR Mapping Ayman F. Habib 27 Stability Analysis: Problem Statement • Some mapping activities utilize imaging systems not intended for mapping applications. • For these systems, we need to investigate whether the estimated interior orientation parameters from temporal calibration sessions are compatible or not. • Statistical testing focuses on testing the changes in the numerical values of the IOP without investigating their impact on the reconstructed object space. • We should develop procedures that investigate the impact of changes in the IOP on the reconstructed object space: Bundle similarity for stability analysis. Photogrammetric & LiDAR Mapping Ayman F. Habib 28 cI cII P.C.I P.C.II Original Image Grid Points Distortion-free Grid Points using IOPI Bundle I Bundle II Distortion-free Grid Points using IOPII Side View Top View Stability Analysis: New Approach ≡ ? Photogrammetric & LiDAR Mapping Ayman F. Habib 29 • The degree of similarity between the bundles from two sets of IOP (RMSE offset) is evaluated in terms of average image space deviation (pixel units). • The RMSE offset is compared to expected image noise. • We have different approaches for the RMSE offset estimation that are commensurate with the utilized geo-referencing technique (e.g., indirect, ISO). • The cameras must meet the following specifications to be deemed stable. – RMSE offset (Tier I mapping) < 1 Pixel – RMSE offset (Tier II mapping) < 1.5 Pixels Stability Analysis: New Approach Photogrammetric & LiDAR Mapping Ayman F. Habib GP xb yb zb XG ZG YG Object point (I) 2. IMU body frame 1. Mapping Reference Frame 3. Image Coordinate System x y z )t(r m b )t(R m b b c r c distyy distxx ypi xpi c i r c ii rS b c R m I r Image point (i) c i b c m bi b c m b m b m I rRtRSrtRtrr )()()( 30 Mounting Parameters Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 31 xb yb zb Xm Object point I )t(r m b )t(R m b b c r b c R 1. Mapping reference frame Ym Zm 2. IMU body frame m I r m I r x. y z Image point i 3. Camera Coordinate System c i r c distyy distxx ypi xpi c iirS GPS/INS-assisted Photogrammetric System c i b c m bi b c m b m b m I rRtRSrtRtrr )()()( System Calibration Photogrammetric System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib Conjugate Points a´ 32 a Object Point (A) • Conceptual Basis: If the system is properly calibrated and geo-referenced, conjugate light rays should intersect at the true position of the corresponding object point (regardless of the flight direction/configuration). Photogrammetric System Calibration System Calibration: IOP + Mounting Parameters Photogrammetric & LiDAR Mapping Ayman F. Habib 33 I. Check whether inaccurate/biased parameters would lead to Y-parallax between conjugate light rays from directly geo-referenced stereo-imagery II. Check whether inaccurate/biased parameters would lead to biases in the derived object points, whose magnitudes and directions depend on the flight configuration III. If some of the system parameters under investigation do not introduce Y- parallax between conjugate light rays or discrepancies between derived points from overlapping imagery in a given flight configuration, control points will be required to derive such parameters System Calibration: Mounting and camera parameters Photogrammetric System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 34 I. Y-Parallax II. Discrepancies III. Control Information Reconstructed Object Space Flightline 1 Reconstructed Object Space Flightline 2 Reconstructed Object Space Control Surface System Calibration: Mounting and camera parameters • Analysis Concept: Photogrammetric System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib • To investigate whether biases in the system mounting parameters will introduce artificial parallax, we generate a pair of normalized images. 35 Photogrammetric System Calibration lb x lb y lb z lx ly lz rx ry rz ln x ln y ln z rn x rn yrnz IMU body frame Left Image Right Image Normalized image pair cncn cc OrOl rb x rb y rb z IMU body frame Object point (I) ln i x ln i y rn i y rn i x • Conjugate points in the normalized images will not have y-parallax. • For the analysis, we introduce systematic errors in the system calibration parameters and investigate whether they will lead to x- or y-parallax in the normalized image coordinates. Photogrammetric & LiDAR Mapping Ayman F. Habib 36 ni pi x rn y rn y ln y rn x ln x ni pi x 1 2 3 4 6 7 1 2 3 4 6 7 Bias in the xp component of the camera parameters Impact of Camera Parameters Errors on the Normalized Images Photogrammetric System Calibration No artificial parallax neither in the X-direction nor in the Y-direction Such finding reveals the fact that it is not possible to estimate the bias in the xp component of the camera parameters from a control-free stereo pair. Photogrammetric & LiDAR Mapping Ayman F. Habib 37 Bias in the yp component of the camera parameters Impact of Camera Parameters Errors on the Normalized Images Photogrammetric System Calibration ni pi y ni pi y rn y ln y rn x ln x 1 2 3 4 6 7 1 2 3 4 6 7 No artificial parallax neither in the X-direction nor in the Y-direction Such finding reveals the fact that it is not possible to estimate the bias in the yp component of the camera parameters from a control-free stereo pair. Photogrammetric & LiDAR Mapping Ayman F. Habib 38 rn y ln y rn x ln x c c y nni l nii c c y nni r nii c c x nni l nii c c x nni r nii 1 2 3 4 6 7 1 2 3 4 6 7 artificial parallax in the X-direction Such finding reveals the fact that it is not possible to estimate the bias in the camera’s principal distance from a control-free stereo pair. Bias in the principal distance of the camera Impact of Camera Parameters Errors on the Normalized Images Photogrammetric System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 1 2 5 6 3 4 1 2 5 6 3 4 No artificial parallax neither in the X-direction nor in the Y-direction Bias in the X component of the lever-arm offset vector 39 Impact of Mounting Parameters Errors on the Normalized Images Such finding reveals the fact that it is not possible to estimate the bias in the X component of the lever-arm offset vector from a control-free stereo pair. Photogrammetric System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 40 Impact of Mounting Parameters Errors on the Normalized Images Bias in the Y component of the lever-arm offset vector 1 2 5 6 3 4 1 2 5 6 3 4 No artificial parallax neither in the X-direction nor in the Y-direction Such finding reveals the fact that itis not possible to estimate the bias in the Y component of the lever-arm offset vector from a control-free stereo pair. Photogrammetric System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 1 2 5 6 3 4 1 2 5 6 3 4 41 artificial parallax in the X-direction Impact of Mounting Parameters Errors on the Normalized Images Bias in the vertical component of the lever-arm offset vector Such finding reveals the fact that it is not possible to estimate the bias in the vertical component of the lever-arm offset vector from a control-free stereo pair. Photogrammetric System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 1 2 5 6 3 4 1 2 5 6 3 4 42 artificial parallax in the X-direction Bias in the boresight roll angle Impact of Mounting Parameters Errors on the Normalized Images Such finding reveals the fact that it is not possible to estimate the bias in the boresight roll angle from a control-free stereo pair. Photogrammetric System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 1 2 5 6 3 4 1 2 5 6 3 4 43 artificial parallax in the X-direction and in the Y-direction Bias in the boresight pitch angle Such finding reveals the possibility of estimating the bias in the boresight pitch angle using a control-free stereo pair. Impact of Mounting Parameters Errors on the Normalized Images Photogrammetric System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 1 2 5 6 3 4 1 2 5 6 3 4 44 artificial parallax in the Y-direction Bias in the boresight yaw angle Impact of Mounting Parameters Errors on the Normalized Images Such finding reveals the possibility of estimating the bias in the boresight yaw angle using a control-free stereo pair. Photogrammetric System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 45 45 Lever-Arm Offset Bias (δΔX) Lever-Arm Offset Bias (δΔY) Lever-Arm Offset Bias (δΔZ) Impact of Mounting Parameters Errors on Reconstructed Object Space Photogrammetric System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib DX DY DZ 46 0 100 200 300 350 -500 -300 -100 100 300 500 -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 X axis (m)Y axis (m) D if f. X ( m ) Forward Backward 0 100 200 300 350 -500 -300 -100 100 300 500 -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 D if f. Y ( m ) X axis (m)Y axis (m) Forward Backward 0 100 200 300 350 -500 -300 -100 100 300 500 -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 D if f. Z ( m ) X axis (m)Y axis (m) Forward Backward Differences between the bias-contaminated and true object-space in the X , Y, and Z directions, after the introduction of a bias in the xp component of the camera parameters (1000m flight height). Such finding reveals the possibility of estimating the bias in the xp of the camera parameters by having flight lines in opposite directions. Photogrammetric & LiDAR Mapping Ayman F. Habib DX DY DZ 47 -200 0 200 400 600 -500 0 500 -1.5 -1 -0.5 0 0.5 1 1.5 X axis (m)Y axis (m) D if f. X ( m ) Forward Backward -200 0 200 400 600 -500 0 500 -1.5 -1 -0.5 0 0.5 1 1.5 D if f. Y ( m ) X axis (m)Y axis (m) Forward Backward -200 0 200 400 600 -500 0 500 -1.5 -1 -0.5 0 0.5 1 1.5 D if f. Z ( m ) X axis (m)Y axis (m) Forward Backward Differences between the bias-contaminated and true object-space in the X , Y, and Z directions, after the introduction of a bias in the xp component of the camera parameters (1800m flight height). The impact of the bias in the xp of the camera parameters is flying-height dependent. Photogrammetric & LiDAR Mapping Ayman F. Habib DX DY DZ 48 0 100 200 300 350 -500 -300 -100 100 300 500 -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 D if f. X ( m ) X axis (m)Y axis (m) Forward Backward 0 100 200 300 350 -500 -300 -100 100 300 500 -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 X axis (m)Y axis (m) D if f. Y ( m ) Forward Backward 0 100 200 300 350 -500 -300 -100 100 300 500 -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 D if f. Z ( m ) X axis (m)Y axis (m) Forward Backward Such finding reveals the possibility of estimating the bias in the yp of the camera parameters by having flight lines in opposite directions. Differences between the bias-contaminated and true object-space in the X , Y, and Z directions, after the introduction of a bias in the yp component of the camera parameters (1000m flight height). Photogrammetric & LiDAR Mapping Ayman F. Habib DX DY DZ 49 -200 0 200 400 600 -500 0 500 -1.5 -1 -0.5 0 0.5 1 1.5 D if f. X ( m ) X axis (m)Y axis (m) Forward Backward -200 0 200 400 600 -500 0 500 -1.5 -1 -0.5 0 0.5 1 1.5 X axis (m)Y axis (m) D if f. Y ( m ) Forward Backward -200 0 200 400 600 -500 0 500 -1.5 -1 -0.5 0 0.5 1 1.5 D if f. Z ( m ) X axis (m)Y axis (m) Forward Backward Differences between the bias-contaminated and true object-space in the X , Y, and Z directions, after the introduction of a bias in the yp component of the camera parameters (1800m flight height). The impact of the bias in the yp of the camera parameters is flying-height dependent. Photogrammetric & LiDAR Mapping Ayman F. Habib DX DY DZ 50 0 100 200 300 350 -500 -300 -100 100 300 500 -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 D if f. X ( m ) X axis (m)Y axis (m) Forward Backward 0 100 200 300 350 -500 -300 -100 100 300 500 -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 D if f. Y ( m ) X axis (m)Y axis (m) Forward Backward 0 100 200 300 350 -500 -300 -100 100 300 500 -1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1 X axis (m)Y axis (m) D if f. Z ( m ) Forward Backward Differences between the bias-contaminated and true object-space in the X , Y, and Z directions, after the introduction of a bias in the camera’s principal distance (1000m fling height) Such finding reveals the fact that it is not possible to estimate the bias in camera’s principal distance by having flight lines in opposite directions. Photogrammetric & LiDAR Mapping Ayman F. Habib DX DY DZ 51 Differences between the bias-contaminated and true object-space in the X , Y, and Z directions, after the introduction of a bias in the camera’s principal distance (1800m fling height) -200 0 200 400 600 -500 0 500 -1.5 -1 -0.5 0 0.5 1 1.5 D if f. X ( m ) X axis (m)Y axis (m) Forward Backward -200 0 200 400 600 -500 0 500 -1.5 -1 -0.5 0 0.5 1 1.5 D if f. Y ( m ) X axis (m)Y axis (m) Forward Backward -200 0 200 400 600 -500 0 500 -1.5 -1 -0.5 0 0.5 1 1.5 D if f. Z ( m ) X axis (m)Y axis (m) Forward Backward The impact of the bias in the camera’s principal distance is flying-height dependent. Photogrammetric & LiDAR Mapping Ayman F. Habib 52 0 100 200 300 -500 -300 -100 100 300 500 -0.1 -0.05 0 0.05 0.1 0.15 X axis (m)Y axis (m) D iff . X ( m ) Forward Backward 0 100 200 300 -500 -300 -100 100 300 500 -0.1 -0.05 0 0.05 0.1 0.15 D iff . Y ( m ) X axis (m)Y axis (m) Forward Backward 0 100 200 300 -500 -300 -100 100 300 500 -0.1 -0.05 0 0.05 0.1 0.15 D iff . Z ( m ) X axis (m)Y axis (m) Forward Backward Differences between the bias-contaminated and true object-space in the X , Y, and Z directions, after the introduction of a bias in the X component of the lever- arm offset (1000m fling height). DX DY DZ Such finding reveals the possibility of estimating the bias in the X component of the lever-arm offset by having flight lines in opposite directions. Photogrammetric & LiDAR Mapping Ayman F.Habib DZ DX DY 53 -200 0 200 400 600 -500 0 500 -0.1 -0.05 0 0.05 0.1 0.15 D if f. X ( m ) X axis (m)Y axis (m) Forward Backward -200 0 200 400 600 -500 0 500 -0.1 -0.05 0 0.05 0.1 0.15 D if f. Y ( m ) X axis (m)Y axis (m) Forward Backward -200 0 200 400 600 -500 0 500 -0.1 -0.05 0 0.05 0.1 0.15 D if f. Z ( m ) X axis (m)Y axis (m) Forward Backward Differences between the bias-contaminated and true object-space in the X , Y, and Z directions, after the introduction of a bias in the X component of the lever- arm offset (1800m fling height). The impact of the bias in the X component of the lever arm offset is flying-height independent. Photogrammetric & LiDAR Mapping Ayman F. Habib 54 0 100 200 300 -500 -300 -100 100 300 500 -0.1 -0.05 0 0.05 0.1 0.15 D iff . X ( m ) X axis (m)Y axis (m) Forward Backward 0 100 200 300 -500 -300 -100 100 300 500 -0.1 -0.05 0 0.05 0.1 0.15 D iff . Z ( m ) X axis (m)Y axis (m) Forward Backward DX DY DZ Differences between the bias-contaminated and true object-space in the X , Y, and Z directions, after the introduction of a bias in the Y component of the lever- arm offset (1000m fling height). Such finding reveals the possibility of estimating the bias in the Y component of the lever-arm offset by having flight lines in opposite directions. 0 100 200 300 -500 -300 -100 100 300 500 -0.1 -0.05 0 0.05 0.1 0.15 X axis (m)Y axis (m) D iff . Y (m ) Forward Backward Photogrammetric & LiDAR Mapping Ayman F. Habib DX DY DZ 55 -200 0 200 400 600 -500 0 500 -0.1 -0.05 0 0.05 0.1 0.15 D if f. X ( m ) X axis (m)Y axis (m) Forward Backward -200 0 200 400 600 -500 0 500 -0.1 -0.05 0 0.05 0.1 0.15 D if f. Y ( m ) X axis (m)Y axis (m) Forward Backward -200 0 200 400 600 -500 0 500 -0.1 -0.05 0 0.05 0.1 0.15 D if f. Z ( m ) X axis (m)Y axis (m) Forward Backward Differences between the bias-contaminated and true object-space in the X , Y, and Z directions, after the introduction of a bias in the Y component of the lever- arm offset (1800m fling height). The impact of the bias in the Y component of the lever arm offset is flying-height independent. Photogrammetric & LiDAR Mapping Ayman F. Habib 56 0 100 200 300 -500 -300 -100 100 300 500 -0.1 -0.05 0 0.05 0.1 0.15 D iff . X ( m ) X axis (m)Y axis (m) Forward Backward 0 100 200 300 -500 -300 -100 100 300 500 -0.1 -0.05 0 0.05 0.1 0.15 D iff . Y ( m ) X axis (m)Y axis (m) Forward Backward 0 100 200 300 -500 -300 -100 100 300 500 -0.1 -0.05 0 0.05 0.1 0.15 D iff . Z ( m ) Y axis (m) X axis (m) Forward Backward Differences between the bias-contaminated and true object-space in the X , Y, and Z directions, after the introduction of a bias in the Z component of the lever- arm offset (1000m fling height). DX DY DZ Such finding reveals the fact that it is not possible to estimate the bias in the Z component of the lever- arm offset by having flight lines in opposite directions. Photogrammetric & LiDAR Mapping Ayman F. Habib DX DY DZ 57 -200 0 200 400 600 -500 0 500 -0.1 -0.05 0 0.05 0.1 0.15 D if f. X ( m ) X axis (m)Y axis (m) Forward Backward -200 0 200 400 600 -500 0 500 -0.1 -0.05 0 0.05 0.1 0.15 D if f. Y ( m ) X axis (m)Y axis (m) Forward Backward -200 0 200 400 600 -500 0 500 -0.1 -0.05 0 0.05 0.1 0.15 D if f. Z ( m ) X axis (m)Y axis (m) Forward Backward Differences between the bias-contaminated and true object-space in the X , Y, and Z directions, after the introduction of a bias in the Z component of the lever- arm offset (1800m fling height). The impact of the bias in the Z component of the lever arm offset is flying-height independent. Photogrammetric & LiDAR Mapping Ayman F. Habib 58 0 100 200 300 -500 -300 -100 100 300 500 -0.1 -0.05 0 0.05 0.1 0.15 D if f. X ( m ) X axis (m)Y axis (m) Forward Backward 0 100 200 300 350 -500 -300 -100 100 300 500 -3 -2 -1 0 1 2 3 X axis (m)Y axis (m) D if f. Y ( m ) Forward Backward 0 100 200 300 350 -500 -300 -100 100 300 500 -2 -1 0 1 2 X axis (m) Y axis (m) D if f. Z ( m ) Forward BackwardDX DY DZ Differences between the bias-contaminated and true object-space in the X , Y, and Z directions, after the introduction of a bias in the boresight roll angle. Such finding reveals the possibility of estimating the bias in the boresight roll angle by having flight lines in opposite directions. Photogrammetric & LiDAR Mapping Ayman F. Habib 59 0 100 200 300 350 -500 -300 -100 100 300 500 -4 -2 0 2 4 X axis (m)Y axis (m) D if f. X ( m ) Forward Backward 0 100 200 300 350 -500 -300 -100 100 300 500 -4 -2 0 2 4 X axis (m)Y axis (m) D if f. X ( m ) Forward Backward 0 100 200 300 350 -500 -300 -100 100 300 500 -0.1 -0.05 0 0.05 0.1 0.15 X axis (m) Y axis (m) D if f. Y ( m ) Forward Backward 0 100 200 300 350 -500 -300 -100 100 300 500 -1 -0.5 0 0.5 1 X axis (m) Y axis (m) D if f. Z ( m ) Forward Backward Differences between the bias-contaminated and true object-space in the X , Y, and Z directions, after the introduction of a bias in the boresight pitch angle. DX DY DZ Such finding reveals the possibility of estimating the bias in the boresight pitch angle by having flight lines in opposite directions. Photogrammetric & LiDAR Mapping Ayman F. Habib 60 0 50 100 150 200 250 300 350 -500 -300 -100 100 300 500 -2 -1 0 1 2 X axis (m)Y axis (m) D if f. X ( m ) Forward Backward 0 100 200 300 350 -500 -300 -100 100 300 500 -1 -0.5 0 0.5 1 X axis (m)Y axis (m) D if f. Y ( m ) Forward Backward 0 100 200 300 -500 -300 -100 100 300 500 -0.1 -0.05 0 0.05 0.1 0.15 D if f. Z ( m ) X axis (m)Y axis (m) Forward Backward DX DY DZ Such finding reveals the fact that flight lines in opposite directions would not lead to additional information for the estimation of the bias in the boresight yaw angle. Differences between the bias-contaminated and true object-space in the X , Y, and Z directions, after the introduction of a bias in the boresight yaw angle. However, parallel flight lines with some side lap would lead to additional information for the estimation of the bias in the boresight yaw angle. Photogrammetric & LiDAR Mapping Ayman F. Habib 350 400 450 500 550 600 650 -500 -300 -100 100 300 500 700 9001000 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 X axis (m) Y axis (m) D if f. X ( m ) Flightline 1 Flightline 2 350 400 450 500 550 600 650 -500 -300 -100 100 300 500 700 900 1000-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 Y axis (m) X axis (m) D if f. Y ( m ) Flightline 1 Flightline 2 350 400 450 500 550 600 650 -500 -300 -100 100 300 500 700 900 1000-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 Y axis (m) X axis (m) D if f. Z ( m ) Flightline 1 Flightline 2 Differences between the bias-contaminated and true object-space in the X , Y, and Z directions, after the introduction of a bias in the boresight yaw angle. Parallel flight lines with some side lap would lead to additional information for the estimation of the bias in the boresight yaw angle. 61 DX DY DZ Photogrammetric & LiDAR Mapping Ayman F. Habib 62 Parameter y-parallax Discrepancies: Flying Direction/ Flying Height/ Point Coord. Dependent Control Requirement planimetric lever arm offset components No Yes/No/No No verticallever arm offset component No No/No/No Yes boresight roll No Yes/Yes/Yes No boresight Pitch Yes Yes/Yes/Yes No boresight yaw Yes No/Yes/Yes No principal point coordinates No Yes/Yes/No No principal distance No No/Yes/No No Summary Photogrammetric System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 63 Parameter Y-parallax Discrepancies: Flying Direction/ Flying Height Dependent Control Requirement Optimum Configuration Planimetric Lever-arm components No Yes/No No Opposite flight lines 100% overlap Vertical Lever-arm component No No/No Yes 1 vertical GCP Boresight Pitch Yes Yes/Yes No Opposite flight lines 100% overlap Boresight yaw Yes Yes/Yes No Parallel flight lines with minimum side lap Boresight roll No Yes/Yes No Opposite flight lines 100% overlap Principal Point Coordinates No Yes/Yes No Opposite flight lines 100% overlap Principal Distance No No/Yes No Flight lines at diff. flying height Summary Photogrammetric System Calibration At Different Flying Heights Photogrammetric & LiDAR Mapping Ayman F. Habib Optimum Flight Configuration for the Estimation of the Mounting Parameters 64 Photogrammetric System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 65 Optimum Flight Configuration for the System Calibration (Mounting Parameters, xp, yp, c) Photogrammetric System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib • So far, we assumed that we are dealing with a single-camera system. • For each image, we have a set of exterior orientation parameters. • Modern mobile mapping systems can utilize multi-camera imaging system. • The direct incorporation of the GPS/INS-derived position and orientation information in the collinearity equations is the optimum way of dealing with such systems. – Modified collinearity equations 66 Photogrammetric System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 67 Airborne Mobile Mapping System Lever arm and boresight angles between the different cameras Mounting Parameters Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 68 Lever arm and boresight angles between the different cameras Terrestrial Mobile Mapping System Mounting Parameters Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 69 Lever arm and boresight angles between the different cameras Terrestrial Mobile Mapping System Mounting Parameters Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 70 For Multi-Camera System Mounting Parameters Calibration x y z x y z x y z x y z x z y x b y b z b Camera “0” Camera “1” Camera “2” Camera “3” Camera “j” IMU Body Frame X G Y G Z G • Lever arm and boresight angles between the different cameras • Lever arm and boresight angles between the IMU body frame and the cameras •These parameters are not independent. Photogrammetric & LiDAR Mapping Ayman F. Habib 71 XG x y z x y z x y z x y z x z y Camera “0” Camera “1” Camera “2” Camera “3” Camera “j” YG ZG )(0 tR m c )(0 tr m c )(tR m cj )(tr m cj 0 1 c c R 0 1 c c r 0c cjR0c cjr # Unknowns = N_cam * N_epoch * 6 For Multi-Camera System Mounting Parameters Calibration Indirect Geo-Referencing for Multi-Camera Systems with Additional ROC Photogrammetric & LiDAR Mapping Ayman F. Habib 72 • If we have N_cam cameras capturing imagery for N_epoch: )()()()( )()()()( )()()()( )()()()( )()()()( )()()()( )()()()( )()()()( )()()()( __11 __11 __11 3311 3311 3311 2211 2211 2211 _ 1 _ 11 _ 3 1 3 11 3 2 1 2 11 2 _ 1 _ 11 _ 3 1 3 11 3 2 1 2 11 2 _ 1 _ 11 _ 3 1 3 11 3 2 1 2 11 2 epochN m cepochN c m m c c m c c epochN m cepochN c m m c c m c c epochN m cepochN c m m c c m c c m c c m m c c m c c m c c m m c c m c c m c c m m c c m c c m c c m m c c m c c m c c m m c c m c c m c c m m c c m c c tRtRtRtRR tRtRtRtRR tRtRtRtRR tRtRtRtRR tRtRtRtRR tRtRtRtRR tRtRtRtRR tRtRtRtRR tRtRtRtRR camNcamNcamN camNcamNcamN camNcamNcamN ))()(()())()(()( ))()(()())()(()( ))()(()())()(()( ))()(()())()(()( ))()(()())()(()( ))()(()())()(()( ))()(()())()(()( ))()(()())()(()( ))()(()())()(()( ___111 ___111 ___111 333111 333111 333111 222111 222111 222111 1_ 1 1_ 11 _ 13 1 13 11 3 12 1 12 11 2 1_ 1 1_ 11 _ 13 1 13 11 3 12 1 12 11 2 1_ 1 1_ 11 _ 13 1 13 11 3 12 1 12 11 2 epochN m cepochN m cepochN c m m c m c c m c c epochN m cepochN m cepochN c m m c m c c m c c epochN m cepochN m cepochN c m m c m c c m c c m c m c c m m c m c c m c c m c m c c m m c m c c m c c m c m c c m m c m c c m c c m c m c c m m c m c c m c c m c m c c m m c m c c m c c m c m c c m m c m c c m c c trtrtRtrtrtRr trtrtRtrtrtRr trtrtRtrtrtRr trtrtRtrtrtRr trtrtRtrtrtRr trtrtRtrtrtRr trtrtRtrtrtRr trtrtRtrtrtRr trtrtRtrtrtRr camNcamNcamN camNcamNcamN camNcamNcamN For Multi-Camera System Mounting Parameters Calibration Indirect Geo-Referencing for Multi-Camera Systems with Additional ROC Relative relationship between the different cameras and the reference one are constant at epochs t1, t2, …, tN_epoch Total Number of Constraints: 6 (N_epoch – 1) (N_cam -1) • Total number of EOP: 6 * N_cam* N_epoch • Total number of constraints: 6 * (N_cam – 1) * (N_epoch – 1) • Total number of independent EOP: 6 * N_epoch + 6 * (N_cam – 1) Photogrammetric & LiDAR Mapping Ayman F. Habib x y z x y z x y z x y z x z y Camera “0” Camera “1” Camera “2” Camera “3” Camera “j” XG YG ZG )(0 tR m c )(0 tr m c )(tR m cj )(tr m cj xb yb zb IMU Body Frame b c r 0 b c R 0 b cj r b cj R )(tR m b )(tr m b 73 For Multi-Camera System Traditional ISO (collinearity equations + GPS/INS observations) Mounting Parameters Calibration # of Add. Obs. = N_epoch * N_cam * 6 (Some Repeated Observations) # Unknowns = N_cam * N_epoch * 6 + N_cam * 6 camNjrtRtrjtr cj b m cj m cj m b _1:)()(}){( camNjRtRjtR cj b m cj m b _1:),,()(}{)( kfw j j j j c i m c c i m c m I rtRStrr )()( Photogrammetric & LiDAR Mapping Ayman F. Habib 74 GP xb yb zb XG ZG YG Object point (I) 2. IMU body frame 1. Mapping Reference Frame 3. Image Coordinate System x y z )t(r m b )t(R m b b c r c distyy distxx ypi xpi c i r c ii rS b c R m I r Image point (i) c i b c m bi b c m b m b m I rRtRSrtRtrr )()()( Photogrammetric Positioning: New Model New ISO Model (Direct Incorporation of GPS/INS observations): Single-Camera System )()()( / t e e e t Z Y X t Z Y X m bZ Y X m b INSGPS m b )()()( / t e e e tt m b m b INSGPS m b k w k w k w Mounting Parameters Photogrammetric & LiDAR Mapping Ayman F. Habib x y z x y z x y z x y z x z y Camera “0” Camera “1” Camera “2” Camera “3” Camera “j” Xm Ym Zm xb yb zb IMU Body Frame b c r 0 b c R 0 )t(r m b )t(R m b b cj r b cj R )( )()( j j j j c i b c m b c i b c m b m b m I rRtRSrtRtrr 75 For Multi-Camera System Mounting Parameters Calibration New ISO Model (Direct Incorporation of GPS/INS observations ):Multi-Camera System )()()( / t e e e t Z Y X t Z Y X m bZ Y X m b INSGPS m b )()()( / t e e e tt m b m b INSGPS m b k w k w k w # of Add. Obs. = N_epoch * 6 # Unknowns = N_epoch * 6 + N_cam * 6 Photogrammetric & LiDAR Mapping Ayman F. Habib 76 Mounting Parameters Calibration x y z x y z x y z x y z x z y Camera “0” Camera “1” Camera “2” Camera “3” Camera “j” XG YG ZG )(0 tR m c )(0 tr m c 0 1 c c R 0 1 c c r 0c cj R0c cj r For Multi-Camera System Indirect Geo-Referencing with Built-in Relative Orientation Constraints xb yb zlb Virtual IMU Body Frame j j j j c i b c m b c i b c m b m b m I rRtRSrtRtrr )()()( )()()( / t e e e t Z Y X t Z Y X m bZ Y X m b INSGPS m b )()()( / t e e e tt m b m b INSGPS m b k w k w k w 0 0 0 b cr Z Y X 0 0 0 b cr k w Do not exist Do not exist Treated as constant )( )()( jr jr jr jrr c i c c m c c i c c m c m c m I rRtRSrtRtrr Photogrammetric & LiDAR Mapping Ayman F. Habib 77 For Multi-Camera System General Model: Incorporation of prior information on the ROP among the cameras x y z x y z x y z x y z x z y Camera “0” Camera “1” Camera “2” Camera “3” Camera “j” 0 1 c c r 0 1 c c R 0c cj R xb yb zb IMU Body Frame b c R 0 XG YG ZG )(tr m b b c r 0 0c cj r )t(R m b )( )( )()( jr jr jr jrr c i c c b c m b c i c c b c m b b c m b m b m I rRRtRSrRtRrtRtrr )()()( / t e e e t Z Y X t Z Y X m bZ Y X m b INSGPS m b )()()( / t e e e tt m b m b INSGPS m b k w k w k w cr cjZ Y X cr cj cr cj e e e Z Y X )prior( Z Y X cr cj cr cj cr cj e e e )prior( k w k w k w b crZ Y X b cr b cr e e e Z Y X )prior( Z Y X b cr b cr b cr e e e )prior( k w k w k w Photogrammetric Positioning: General Model Photogrammetric & LiDAR Mapping Ayman F. Habib x y z x y z x y z x y z x z y Camera “0” Camera “1” Camera “2” Camera “3” Camera “j” XG YG ZG )(0 tR m c )(0 tr m c 0 1 c c R 0 1 c c r 0c cj R0c cj r 78 For Multi-Camera System Special case of the general model: Indirect Geo-Referencing with Built-in Relative Orientation Constraints )( )()( jr jr jr jrr c i c c m c c i c c m c m c m I rRtRSrtRtrr Mounting Parameters Calibration xb yb zlb Virtual IMU Body Frame )( )( )()( jr jr jr jrr c i c c b c m b c i c c b c m b b c m b m b m I rRRtRSrRtRrtRtrr General Model )()()( / t e e e t Z Y X t Z Y X m bZ Y X m b INSGPS m b )()()( / t e e e tt m b m b INSGPS m b k w k w k w cr cjZ Y X cr cj cr cj e e e Z Y X )prior( Z Y X cr cj cr cj cr cj e e e )prior( k w k w k w 0 0 0 )prior( Z Y X b cr 0 0 0 )prior( b cr k w Treated as constant Do not exist Do not exist Might exist Might exist Photogrammetric & LiDAR Mapping Ayman F. Habib x y z x y z x y z x y z x z y Camera “0” Camera “1” Camera “2” Camera “3” Camera “j” XG YG ZG xb yb zb IMU Body Frame b c r 0 b c R 0 )(tr m b )(tR m b b cj r b cj R 79 For Multi-Camera System Special case of the general model: ISO Model without ROP Prior Information )( )()( j j j j c i b c m b c i b c m b m b m I rRtRSrtRtrr Mounting Parameters Calibration )( )( )()( jr jr jr jrr c i c c b c m b c i c c b c m b b c m b m b m I rRRtRSrRtRrtRtrr General Model x y z Virtual Reference Camera cr cjZ Y X cr cj cr cj e e e Z Y X )prior( Z Y X cr cj cr cj cr cj e e e )prior( k w k w k w )()()( / t e e e t Z Y X t Z Y X m bZ Y X m b INSGPS m b )()()( / t e e e tt m b m b INSGPS m b k w k w k w 0 0 0 )prior( Z Y X b cr 0 0 0 )prior( b cr k w Treated as constant Does not exist Does not exist Photogrammetric & LiDAR Mapping Ayman F. Habib 80 • Photogrammetric reconstruction is based on redundant measurements. • Results from the photogrammetric triangulation gives quantitative measures of the precision of the reconstruction outcome. – Variance component (overall measure of the quality of fit between the observed quantities and the used model) – Variance-covariance matrix for the derived object coordinates – These values can be compared with expected nominal values • Independent measure for accuracy verification can be established using check point analysis. – Photogrammetric coordinates are compared with independently measured coordinates (e.g., GPS survey) RMSE analysis. Photogrammetric Quality Control Photogrammetric & LiDAR Mapping Ayman F. Habib 81 • Precision of a single model: If we have – Bundle block adjustment with additional parameters that compensate for various distortions – Regular blocks with 60% overlap and 20% side lap – Signalized targets space image in thegiven are valuesprecision These cameras)(SWA distance principalcameratheof%004.0 cameras) WA and(NA distance principalcameratheof%003.0 3 Z Z XY m Precision of Photogrammetric Reconstruction Photogrammetric & LiDAR Mapping Ayman F. Habib 82 Check Point Analysis Accuracy of Photogrammetric Reconstruction Photogrammetric & LiDAR Mapping Ayman F. Habib 83 40 m 40 m 15 m 10 m10 m10 m10 m 10 m 10 m 10 m 10 m 10 m 10 m 15 m 1 2 3 456 7 8 9 10 11 12 wall w al l w al l wall YG ZG Results from Simulated Data Configuration used for the estimation of the ROP among the camerasMounting Parameters Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 84 400 m 400 m 150 m 100 m100 m100 m100 m 100 m 100 m 100 m 100 m 100 m 100 m 150 m 1 2 3 456 7 8 9 10 11 12 wall w al l w al l wall YG ZG Results from Simulated Data Mounting Parameters Calibration Configuration used for the ISO with/without prior information on the ROP Photogrammetric & LiDAR Mapping Ayman F. Habib 85 ROP among the Cameras Reference Camera: Camera “0” Mounting Parameters Calibration Δω (o±″) diff (″) Δφ (o±″) diff (″) Δκ (o±″) diff (″) ΔX (m±m) diff (m) ΔY (m±m) diff (m) ΔZ (m±m) diff (m) “1” 1.00726 ±24.8 26.1 -0.51046 ±25.2 -37.6 -1.99529 ±17.9 17.0 -0.05 ±0.0011 0.00 -1.45 ±0.0024 0.00 0.05 ±0.0028 0.00 “2” -40.99046 ±31.2 34.3 -0.20617 ±34.8 -22.2 -0.98933 ±45.4 38.4 -0.05 ±0.0025 0.00 -1.50 ±0.0036 0.00 0.60 ±0.0040 0.00 “3” -88.99505 ±45.3 17.8 1.97783 ±54.5 -79.8 -0.68205 ±59.8 64.6 -0.05 ±0.0034 0.00 -1.51 ±0.0051 -0.01 1.69 ±0.0052 -0.01 “4” -127.99759 ±41.6 8.7 0.47521 ±84.1 -89.3 -0.39339 ±49.3 23.8 -0.05 ±0.0039 0.00 -1.46 ±0.0063 -0.01 2.44 ±0.0051 -0.01 Photogrammetric & LiDAR Mapping Ayman F. Habib 86 ISO Mounting Parameters Calibration Scenario I: good distribution of the points in the imagery Δω (deg± sec) Diff (sec) Δφ (deg± sec) Diff (sec) Δκ (deg± sec) Diff (sec) ΔX (m± m) Diff. (m) ΔY (m± m) Diff. (m) ΔZ (m± m) Diff. (m) ISO without ROP Prior Information “0” -1.02037 ±50.0 -73.3 -0.52277 ±52.1 -82.0 1.29626 ±31.9 -13.5 0.09 ±0.0578 -0.01 0.51 ±0.0453 0.01 -1.48 ±0.0369 0.07 “1” 1.01157 ±24.7 41.6 -0.50661 ±25.6 -23.8 -2.00714 ±19.5 -25.7 -0.04 ±0.0111 0.01 -1.46 ±0.0111 -0.01 0.03 ±0.0273 -0.02 “2” -40.98329 ±30.6 60.2 -0.19405 ±31.7 21.4 -0.99312 ±38.7 24.8 -0.01 ±0.0212 0.04 -1.53 ±0.0278 -0.03 0.55 ±0.0388 -0.05 “3” -88.96407 ±46.0 129.3 1.99550 ±48.0 -16.2 -0.68499 ±48.7 54.1 0.00 ±0.0318 0.05 -1.59 ±0.0454 -0.09 1.65 ±0.0473 -0.05 “4” -127.98520 ±41.4 53.3 0.49494 ±66.7 -18.2 -0.38421 ±41.8 56.9 0.03 ±0.0349 0.08 -1.49 ±0.0574 -0.04 2.37 ±0.0392 -0.08 Δω (deg± sec) Diff (sec) Δφ (deg± sec) Diff (sec) Δκ (deg± sec) Diff (sec) ΔX (m± m) Diff. (m) ΔY (m± m) Diff. (m) ΔZ (m± m) Diff. (m) ISO with ROP Prior Information “0” -0.99957 ±37.4 1.5 -0.50688 ±28.1 -24.8 1.29701 ±27.9 -10.8 0.16 ±0.0420 0.06 0.45 ±0.0281 -0.05 -1.53 ±0.0287 0.02 “1” 1.00003 ±13.6 0.1 -0.50102 ±13.8 -3.7 -1.99915 ±12.6 3.0 -0.05 ±0.0066 0.00 -1.46 ±0.0067 -0.01 0.05 ±0.0083 0.00 “2” -40.99790 ±13.5 7.6 -0.19356 ±13.8 23.2 -1.00318 ±13.0 -11.5 -0.03 ±0.0076 0.02 -1.50 ±0.0080 0.00 0.61 ±0.0082 0.01 “3” -88.99617 ±14.4 13.8 1.99999 ±14.2 0.0 -0.69842 ±13.9 5.7 -0.05 ±0.0079 0.00 -1.51 ±0.0085 -0.01 1.70 ±0.0082 0.00 “4” -128.00221 ±14.7 -8.0 0.50729 ±15.0 26.3 -0.39904 ±14.0 3.5 -0.04 ±0.0081 0.01 -1.45 ±0.0085 0.00 2.47 ±0.0083 0.02 Slight improvements Photogrammetric & LiDAR Mapping Ayman F. Habib 87 ISO Mounting Parameters Calibration Scenario II: poor distribution of the points in the imagery Δω (deg±sec) Diff.(sec) Δφ (deg±sec) Diff.(sec) Δκ (deg±sec) Diff.(sec) ΔX (m±m) Diff.(m) ΔY (m±m) Diff.(m) ΔZ (m±m) Diff.(m) ISO without ROP Prior Information “0” -1.01261 ±83.6 -45.4 -0.54072 ±79.1 -146.6 1.31125 ±140.7 40.5 -0.01 ±0.0810 -0.11 0.50 ±0.0851 0.00 -1.35 ±0.1327 0.20 “1” 1.01825 ±44.9 65.7 -0.49631 ±46.5 13.3 -1.99811 ±136.7 6.8 0.00 ±0.0335 0.05 -1.47 ±0.0327 -0.02 -0.10 ±0.1049 -0.15 “2” -41.01139 ±131.3 -41.0 -0.19463 ±137.3 19.3 -1.01321 ±165.7 -47.6 0.01 ±0.1257 0.06 -1.34 ±0.2002 0.16 0.36 ±0.1942 -0.24 “3” -88.94687 ±135.5 191.3 2.03304 ±181.9 118.9 -0.67732 ±136.9 81.6 0.15 ±0.1126 0.20 -1.65 ±0.1533 -0.15 1.59 ±0.1923 -0.11 “4” -127.96319 ±126.3 132.5 0.48415 ±182.9 -57.1 -0.38947 ±154.4 37.9 0.08 ±0.1135 0.13 -1.72 ±0.1663 -0.27 2.44 ±0.1849 -0.01 Δω (deg±sec) Diff.(sec) Δφ (deg±sec) Diff.(sec) Δκ (deg±sec) Diff.(sec) ΔX (m± m) Diff.(m) ΔY (m± m) Diff.(m) ΔZ (m± m) Diff.(m) ISO with ROP Prior Information “0” -0.98192 ±45.6 65.1 -0.51156 ±32.5 -41.6 1.30811 ±42.2 29.2 0.10 ±0.0437 0.00 0.41 ±0.0356 -0.09 -1.53 ±0.0481 0.02 “1” 0.99956 ±13.5 -1.6 -0.50046 ±13.5 -1.7 -1.99199 ±15.5 28.8 -0.05 ±0.0073 0.00 -1.46 ±0.0073 -0.01 0.05 ±0.0078 0.00 “2” -40.99696 ±14.4 10.9 -0.19499 ±14.1 18.0 -1.00202 ±15.4 -7.3 -0.03 ±0.0077 0.02 -1.49 ±0.0078 0.01 0.61 ±0.0078 0.01 “3” -89.00064 ±14.8 -2.3 2.00299 ±14.1 10.8 -0.69806 ±15.5 7.0 -0.05 ±0.0076 0.00 -1.50 ±0.0078 0.00 1.70 ±0.0077 0.00 “4” -128.00085 ±15.2 -3.1 0.50711 ±14.9 25.6 -0.39812 ±15.5 6.8 -0.04 ±0.0077 0.01 -1.45 ±0.0078 0.00 2.47 ±0.0078 0.02 0.20 0.02 0.00 0.01 -0.24 0.16 0.01 0.20 -0.15 191.3 118.9 -0.15 -2.3 10.8 0.00 0.00 0.00 -3.1 -0.27 132.5 Significant improvements Photogrammetric & LiDAR Mapping Ayman F. Habib 88 Concluding Remarks • QA/QC procedures are essential for any spatial data acquisition system. • QA/QC are well established for photogrammetric mapping systems. – More attention should be given to medium-format digital imaging systems (specifically, stability analysis). – In-situ calibration of GPS/INS-assisted photogrammetric systems • Camera parameters (principal point coordinates & principal distance) • Mounting parameters – Multi-camera systems require special attention (Relative Orientation Constraint among the cameras) Photogrammetric & LiDAR Mapping Ayman F. Habib Concluding Remarks • The work has introduced a general mathematical model with the following characteristics: – Capable of handling any number of cameras onboard the mapping platform (in the absence or presence of GPS/INS information), and – Allows for the incorporation of prior information about the ROP among the cameras in the ISO. • Special cases can be derived from the general model: – Indirect geo-referencing with built-in Relative Orientation Constraints (ROC) among the cameras – ISO for multi-camera system without prior information about the ROC among the cameras 89 Photogrammetric & LiDAR Mapping Ayman F. Habib • Conducted experiments have demonstrated that: – In the presence of a reasonable imaging geometry and good tying among the images, the use of prior ROP information leads to slight improvements in the accuracy of the estimated parameters. – On the other hand, under the scenario where a poor tying among the images is present, significant improvements are observed (this might happen in terrestrial mobile mapping systems – fewer identifiable features in open areas). • Thus, in addition to the simplification of the mathematical model and the ability to handle several scenarios as special cases, the proposed model has significant advantages when faced with weaker geometric configuration. • Future work will focus on more testing using real datasets from terrestrial and airborne mobile mapping systems. 90 Concluding Remarks Ayman F. Habib MMT 2013 – Summer School(Taiwan) Airborne LiDAR Mapping 91 Laser scanner Photogrammetric & LiDAR Mapping Ayman F. Habib 92 ALS 60 (Leica Geosystems) Operational LiDAR Systems OPTECH ALTM GEMINI long-range RIEGL LMS-Q680i Photogrammetric & LiDAR Mapping Ayman F. Habib 93 LiDAR Principles Three Measurement Systems 1. GNSS 2. IMU 3. Laser scanner emits laser beams with high frequency and collects the reflections. INS GNSS GNSS IMU Direct acquisition of high density and accurate topographic data Photogrammetric & LiDAR Mapping Ayman F. Habib 94 xlb ylb zlb xlu zlu xb yb zb ylu XG ZG YG 2. IMU body frame 1. Mapping Reference Frame Object point (I) Laser beam 3. Laser unit 4. Laser beam )t(r m b )t(R m b b lu r b lu R lu lb R ),( lb I)S,S( lu lb),,( b lu m b)Z,Y,X( b lu m b m b m I rRR)t(Rr)t(R)t(rr kw LiDAR Principles m Ir lbIr Photogrammetric & LiDAR Mapping Ayman F. Habib 95 • QA activities/measures include: – Optimum mission time – Distance to GNSS base station – Flying height – Pulse repetition rate – Beam divergence angle – Scan angle – Percentage of overlap – System calibration LiDAR Quality Assurance Laser scanner Photogrammetric & LiDAR Mapping Ayman F. Habib 96 xlb ylb zlb xlu zlu xb yb zb ylu XG ZG YG 2. IMU body frame 1. Mapping Reference Frame Object point (I) Laser beam 3. Laser unit 4. Laser beam )t(r m b )t(R m b b lu r b lu R lu lb R ),( lb I)S,S( lu lb),,( b lu m b)Z,Y,X( b lu m b m b m I rRR)t(Rr)t(R)t(rr kw LiDAR QA: System Calibration m Ir lbIr Photogrammetric & LiDAR Mapping Ayman F. Habib 97 • Laboratory Calibration (conducted by the system manufacturer) – Calibration of individual system components, – Mirror to IMU misalignment, – Mirror to IMU lever arm, and – Mirror to reference point • Platform Calibration – Reference point to GPS antenna LiDAR QA: System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 98 • In-Flight Calibration (refines the estimated parameters during the laboratory & platform calibration): – Utilizes a calibration test field composed of control surfaces for the estimation of biases in the LiDAR system parameters – The observed discrepancies between the LiDAR and control surfaces are used to determine the biases in the system parameters (e.g., boresighting roll and pitch angles and scale parameters). laser point firing point d LiDAR QA: System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 99 • Status of current calibration methods: – There is lack of a commonly accepted calibration methodology. – System raw measurements are required. – Estimated parameters are limited. – Manual and empirical approaches are utilized. – Calibration sites with control targets are required. • For example, buildings and runways – Calibration is not possible for end-users using point cloud coordinates in overlapping strips. LiDAR QA: System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 100 I. Y-Parallax II. Discrepancies III. Control Information Reconstructed Object Space Flightline 1 Reconstructed Object Space Flightline 2 Reconstructed Object Space Control Surface System Calibration: Mounting and laser scanner parameters • Analysis Concept: LiDAR QA: System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 101 • Conceptual Basis: Estimate the system parameters that minimize discrepancies between derived surfaces from multiple flight lines while reducing ground control requirements • This process requires establishing the optimal flight configuration that maximizes the impact of biases in the system parameters. LiDAR QA: System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib f Opposite directions with 100% overlap ratio Biases in system parameters Lever-arm δΔX Lever-arm δΔY Lever-arm δΔZ Boresight δΔω Boresight δΔφ Boresight δΔκ Range bias δρ Scale bias of S.A. δS X Y w δΔX δΔY δΔω δΔφ YG ZG XG Ground Coordinate System S trip A S trip B Z Y X Ba ck wa rd Fo rw ard 102 LiDAR QA: System Calibration Optimum Flight Configuration Photogrammetric & LiDAR Mapping Ayman F. Habib f Biases in system parameters Lever-arm δΔX Lever-arm δΔY Lever-arm δΔZ Boresight δΔω Boresight δΔφ Boresight δΔκ Range bias δρ Scale bias of S.A. δS Sk δΔκ δΔκ δΔφ δΔφ βT βT δSβT δSβT δρ δρ D x YG ZG XG Ground Coordinate System S trip A S trip BZ Y X Fo rw ard Fo rw ard 103 Same direction with some sidelap LiDAR QA: System Calibration Optimum Flight Configuration Photogrammetric & LiDAR Mapping Ayman F. Habib 104 Parameter Discrepancies: Flying Direction/ Flying Height/ Scan Angle Dependent Control Requirement planimetric lever arm offset components Yes/No/No No vertical lever arm offset component No/No/No Yes boresight roll Yes/Yes/Yes No boresight pitch Yes/Yes/No No boresight yaw No/No/Yes No range bias No/Yes-No*/Yes Yes mirror scan angle scale No/Yes/Yes No Optimum Flight Configuration LiDAR QA: System Calibration * The impact is flying height dependent for conjugate points (small discrepancies are generated among conjugate point) and independent when considering points mapped using the same scan angle. Photogrammetric & LiDAR Mapping Ayman F. Habib 105 Optimum Flight Configuration LiDAR QA: System Calibration Impact of the range bias on strips captured at different flying heights 21 HH XZXZ 21 HH XZXZ H1 H2 Photogrammetric & LiDAR Mapping Ayman F. Habib 106 Parameter Discrepancies: Flying Direction/ Flying Height/ Scan Angle dependent? Control Required? Optimum Requirement Lever arm offset across flight direction ΔX Yes/No/No No Two flight lines in opposite directions with 100% side lap along flight direction ΔY Yes/No/No No Two flight lines in opposite directions with 100% side lap (flying height H1) vertical ΔZ* No/No/No Yes One vertical control point Boresight pitch angle Δω Yes/Yes/No No Two flight lines in opposite directions with 100% side lap (flying height H2) Boresight roll angle Δφ Yes/Yes/Yes No Two flight lines in opposite directions with 100% side lap Boresight yaw angle Δκ No/No/Yes No Two flight lines in the same direction with minimum side lap Range bias Δρ* No/No/Yes Yes One vertical control point Mirror scan angle scale S No/Yes/Yes No Two flight lines in the same direction with minimum side lap *The vertical lever arm offset component and the range bias cannot be simultaneously estimated. Optimum Flight Configuration LiDAR QA: System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 107 Optimum Flight Configuration LiDAR QA: System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib • The DPRG has developed several LiDAR system calibration techniques that are commensurate with the nature of available data. –Simplified Calibration: With some constraints on the flight configuration and ground coverage, we can conduct the calibration using only the point cloud coordinates. –Quasi-Rigorous Calibration: Using the trajectory data and time- tagged point cloud coordinates, we can estimate the system parameters with fewer constraints on the flight configuration. –Rigorous Calibration: With the availability of raw measurements, the calibration can be conducted without any assumptions regarding the flight configuration and ground coverage. 108 LiDAR QA: System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib Overlapping strips Discrepancies 3D Transformation Rotation Shifts CalibrationParameters • LiDAR Data in Overlapping Parallel Strips Point cloud coordinates Raw measurements are not necessarily available Simplified Calibration LiDAR QA: System Calibration 109 • Assumptions: o Linear scanner, o Vertical scanner, o Parallel flight lines, o Terrain-height variations are minimal compared to the flying height, and o Small biases in the boresight angles • Can handle any type of terrain coverage • Cannot handle control points Photogrammetric & LiDAR Mapping Ayman F. Habib Quasi-Rigorous Calibration 110 • LiDAR Data in Overlapping Strips Point cloud coordinates with the time tag Time-tagged trajectory LiDAR QA: System Calibration • Assumptions: o Vertical scanner, o Small biases in the boresight angles • Can handle cross strips • Can handle any type of terrain coverage • Can handle control points Photogrammetric & LiDAR Mapping Ayman F. Habib Rigorous Calibration 111 • LiDAR Data in Overlapping Strips Point cloud coordinates together with the system raw measurements (position and the attitude of each pulse as well as the measured scan angles and ranges) LiDAR QA: System Calibration • Assumptions: o None • Can handle cross strips • Can handle any type of terrain coverage • Can handle control points Photogrammetric & LiDAR Mapping Ayman F. Habib 112 Rigorous Quasi-Rigorous LiDAR Geometric Model Data Requirements Rigorous LiDAR Equation Simplified LiDAR Equation Raw measurements Time-tagged LiDAR point cloud and trajectory position LiDAR QA: System Calibration Photogrammetric & LiDAR Mapping Ayman F. Habib 113 Quasi-Rigorous Approach Assumptions Linear Scanner Vertical LiDAR unit (steady platform) Relatively small boresight angles Simplified LiDAR Equation )Scos( )Ssin( cossin sincos Z Y X cossin sincos )t(rr m b m I w wk k kk kk kk kk 0 1 1 1 100 0 0 100 0 0 ),( lb I)S( lu lb),,( b lu m b)Z,Y,X( b lu m b m b m I rRRRrR)t(rr kwkk )()( Photogrammetric & LiDAR Mapping Ayman F. Habib Trajectory points between (t-Δt) to (t+Δt) Estimated position at time t 114 Quasi-Rigorous Approach z x cossin sincos Z Y X cossin sincos )t(rr m b m I 0 1 1 1 100 0 0 100 0 0 w wk k kk kk kk kk Estimated trajectory point at time t Fitted trajectory line LiDAR point captured at time t x z YG κ β XG ZG Ground Coordinate System Estimated encoder angle (Xt, Yt, Zt) (Xot, Yot, Hot) For each LiDAR point, we compute: – The heading of the trajectory (κ), – The flying height above that point (-z), – The lateral distance (x) between the LiDAR point and the trajectory, and – The mirror scan angle () using (x, z). Photogrammetric & LiDAR Mapping Ayman F. Habib 115 Rigorous Approach No Assumptions Can handle linear and elliptical scanners Can handle steady and unsteady platforms Small/large boresight angles Rigorous LiDAR Equation ),( lb I)S,S( lu lb),,( b lu m b)Z,Y,X( b lu m b m b m I rRR)t(Rr)t(R)t(rr kw Photogrammetric & LiDAR Mapping Ayman F. Habib LSA: Observation Equations (Quasi-rigorous) 116 0XX B True A True For conjugate points in overlapping strips we have: nfG l,xf)True(X nbBiased l,xfX elxxfX nbTrue , 0,, BnbAnb elxxfelxxf Linearization: 0)(/)(/, )(/)(/, ,, ,, BlxlxBnb AlxlxAnb elfxxflxf elfxxflxf BnbB nb AnbAnb ,0~eexAy Photogrammetric & LiDAR Mapping Ayman F. Habib 117 0)(X)(X BGAG TrueTrue For conjugate points in overlapping strips we have: nfG l,xf)(X True noG l,xf)(X Predicted el,xxf)(X noG True 0el,xxfel,xxf BnoAno Linearization: 0)e(l/fxx/fl,xf )e(l/fxx/fl,xf B Bn l, o x B nl,o xBno A An l, o x An l, o xAno LSA: Observation Equations (Rigorous) ,0~eexAy Photogrammetric & LiDAR Mapping Ayman F. Habib 118 Blue: Matched Points Red: Non-matched Points Surface 1 Surface 2 Matching Primitives & Matching Procedure Point/Patch Pairs: Closest Patch Procedure Conditions: • Closest patch (within a threshold) • Point located within the patch Photogrammetric & LiDAR Mapping Ayman F. Habib 119 Closest-patch procedure provides conjugate point/patch pairs that satisfy the following conditions. We will have conjugate point-patch pairs only whenever the TIN patches represent the physical surface. Non-matches • Closest patch • Normal distance < threshold • Point located within the patch Primitives & Matching Procedure Point/Patch Pairs: Closest Patch Procedure Photogrammetric & LiDAR Mapping Ayman F. Habib 120 Surface 1: 44,156 points Surface 2: 45,520 patches Match Primitives & Matching Procedure Point/Patch Pairs: Closest Patch Procedure Photogrammetric & LiDAR Mapping Ayman F. Habib 121 Green: Reference Surface Blue: Matches Red: Non-matches Primitives & Matching Procedure Point/Patch Pairs: Closest Patch Procedure Photogrammetric & LiDAR Mapping Ayman F. Habib 122 Non-matches are typically along edges of buildings and around areas with vegetations Primitives & Matching Procedure Point/Patch Pairs: Closest Patch Procedure Photogrammetric & LiDAR Mapping Ayman F. Habib 123 Conjugate patch to a given point 0 )()( TrueB j TrueA i XX Assuming that A & B are conjugate points Conditions: • Closest patch (within a threshold) • Point located within the patch Primitives & Matching Procedure Point/Patch Pairs: Closest Patch Procedure Photogrammetric & LiDAR Mapping Ayman F. Habib 124 Point-based Observation Equations Modified Least Squares Adjustment d Pseudo-conjugate points CorrectedP Corrected c Corrected b Corrected a exAy Assuming the availability of conjugate points edxAy Modification of the stochastic properties of the random noise vector 0d'P ,~e 0 12 P o Primitives & Matching Procedure Point-Based Adjustment Photogrammetric & LiDAR Mapping Ayman F. Habib 125 • Once the corrections to the system parameters have been estimated, the corrected ground coordinates of the LiDAR points are reconstructed using the estimated parameters: • Quasi-rigorous Approach: • Rigorous Approach: • LiDAR rigorous equation LiDAR Point Cloud Reconstruction ),,,,,,( estestestestestestest SYX kfw estest estestest estestest est estestestest estestestest Biased Biased Biased Corrected Corrected Corrected SxS SzsSx SzcSx x zzcYX zzYX Z Y X Z Y X kkkk kkkk f fkwkkk fkwkkk cos in sinsin cos os sincos sin sin os cossin cos sin sincos ~ ~ ~ ~ ~ ~ Photogrammetric & LiDAR Mapping Ayman F. Habib 126 126 • Quality control is a post-mission procedure to ensure/verify the quality of collected data. • Quality control procedures can be divided into two main categories: – External/absolute QC measures: the LiDAR point cloud is compared with an independently collected surface. • Check point analysis – Internal/relative QC measures:the LiDAR point cloud from different flight lines is compared with each other to ensure data coherence, integrity, and correctness. LiDAR Quality Control Photogrammetric & LiDAR Mapping Ayman F. Habib 127 127 Csanyi, N., Toth, C. (2004). On using LiDAR-specific ground targets. ASPRS Annual Conference, Denver, CO, May 23-28. CD-ROM. EQC: LiDAR Control Targets Photogrammetric & LiDAR Mapping Ayman F. Habib 128 128 Range Data Intensity Data • One should implement a segmentation procedure to derive the LiDAR coordinates of the target. EQC: LiDAR Control Targets Csanyi, N., Toth, C. (2004). On using LiDAR-specific ground targets. ASPRS Annual Conference, Denver, CO, May 23-28. CD-ROM. Photogrammetric & LiDAR Mapping Ayman F. Habib 129 • Surface reconstruction from LiDAR does not have redundancy. – Therefore, we do not have explicit measures in the derived surfaces to assess the quality of LiDAR coordinates. • DPRG Concept: Evaluate the degree of consistency among the LiDAR footprints in overlapping strips. Strip 2 Strip 3 Strip 4 IQC: LiDAR Quality Control Photogrammetric & LiDAR Mapping Ayman F. Habib 130 130 Overlapping strips with common features IQC: LiDAR Quality Control Photogrammetric & LiDAR Mapping Ayman F. Habib 131 131 Check for the presence of biases IQC: LiDAR Quality Control •Derive quantitative estimate of the necessary transformation parameters (shifts & rotations) for the co- alignment of the captured data from different flight lines. • For a well-calibrated system and with accurate navigation information, the transformation parameters should be very close to zero. Photogrammetric & LiDAR Mapping Ayman F. Habib 132 Conditions: • Closest patch • Normal distance < threshold • Point located within the patch 0 1 1 1 1 ccc bbb aaa iii ppp ppp ppp qqq ZYX ZYX ZYX ZYX i i i i i i q q q T T T q q q Z Y X RS Z Y X Z Y X Where: XT, YT, ZT, S, Ω, Φ,Κ ),,,,,,( SZYX TTT Point/Patch Pairs: Iterative Closest Patch (ICPatch) IQC: LiDAR Quality Control Photogrammetric & LiDAR Mapping Ayman F. Habib 133 Surface 1: 44,156 points Surface 2: 45,520 patches Register XT (mm) YT (mm) ZT (mm) S Ω °) Φ °) Κ °) Optimal Para.* 0.000 0.000 0.000 1.000 0.000 0.000 0.000 Estimated -0.660 -0.367 0.007 1.001 -0.017 0.002 0.003 Estimated Variance Component 0.122 Average Normal Distance 0.142 m * Assuming the LiDAR data has no biases Point/Patch Pairs: Iterative Closest Patch (ICPatch) IQC: LiDAR Quality Control Photogrammetric & LiDAR Mapping Ayman F. Habib Strip Number Flying Height Direction 1 1150 m N-S 2 1150 m S-N 3 539 m E-W 4 539 m W-E 5 539 m E-W 6 539 m E-W Strip 1 Strip 2 Strip 3 Strip 4 Strip 5 Strip 6 Strip 1 Strip 2 Strip 3 Strip 4 Strip 5 Strip 6 Overlapping Strips Cases % of Overlap Direction Strips 1&2 80% Opposite directions Strips 3&4 25% Opposite directions Strips 4&5 75% Opposite directions Strips 5&6 50% Same direction 134 LiDAR QA/QC: Experimental Results (I) Dataset Description Photogrammetric & LiDAR Mapping Ayman F. Habib 135 Method δΔX(m) δΔY(m) δΔω(°) δΔφ(°) δΔκ(°) Δρ(m) δS Simplified 0.03 -0.01 -26 -91 -19 0.18 0.000046 Quasi-rigorous -0.01 0.02 -40.2 -90.9 -4.58 0.26 -0.000096 Estimated system biases using the Simplified and the Quasi-rigorous methods LiDAR QA: Experimental Results (I) Photogrammetric & LiDAR Mapping Ayman F. Habib 136 Intensity Image (Before) Intensity Image (After) Qualitative Evaluation LiDAR QC: Experimental Results (I) Photogrammetric & LiDAR Mapping Ayman F. Habib 137 1m1m Original Point Cloud LiDAR QC: Experimental Results (I) Photogrammetric & LiDAR Mapping Ayman F. Habib 138 1m1m Adjusted Point Cloud Relative Accuracy Evaluation LiDAR QC: Experimental Results (I) Photogrammetric & LiDAR Mapping Ayman F. Habib Before Calibration After Calibration Strips 1&2 Strips 1&2 XT (m) YT (m) ZT (m) XT (m) YT (m) ZT (m) 1.10 -0.32 -0.01 0.11 0.07 -0.05 w (deg) φ (deg) κ (deg) w (deg) φ (deg) κ (deg) 0.0001 -0.052 -0.002 0.0012 -0.0016 -0.0051 Strips 3&4 Strips 3&4 XT (m) YT (m) ZT (m) XT (m) YT (m) ZT (m) 0.18 0.41 -0.01 -0.01 -0.01 0.01 w (deg) φ (deg) κ (deg) w (deg) φ (deg) κ (deg) 0.0484 -0.0005 -0.0011 0.0052 0.0008 -0.0045 Compatibility between overlapping strips before and after the calibration procedure 139 Relative Accuracy Evaluation LiDAR QC: Experimental Results (I) Photogrammetric & LiDAR Mapping Ayman F. Habib 140 Before Calibration After Calibration Mean ΔX (m) -0.36 -0.10 Mean ΔY (m) 0.67 0.24 Mean ΔZ (m) -0.05 -0.015 σX (m) 0.40 0.11 σY (m) 0.29 0.06 σZ (m) 0.24 0.13 RMSEX (m) 0.53 0.14 RMSEY (m) 0.72 0.24 RMSEZ (m) 0.25 0.20 RMSETOTAL (m) 0.93 0.35 RMSE analysis of the photogrammetric check points using extracted control planar features from the LiDAR data before and after the calibration procedure Absolute Accuracy Evaluation LiDAR QC: Experimental Results (I) Photogrammetric & LiDAR Mapping Ayman F. Habib 141 Source: http://www.isprs.org/publications/related/semana_geomatica05/front/abstracts/Dimecres9/F01.pdf LiDAR QA/QC: Experimental Results (II) Dataset Description Photogrammetric & LiDAR Mapping Ayman F. Habib 142 Dataset captured by a compact LiDAR system built at EPFL operated from the side of a helicopter LiDAR QA/QC: Experimental Results (II) Dataset Description Photogrammetric & LiDAR Mapping Ayman F. Habib 143 Platform attitude variation Flight line ω (o) min/max φ (o) min/max 1 -3.0 / 4.2 6.4 / 9.1 2 -9.4 / -3.0 -4.0 / 1.6 4 6.8 / 8.7 0.6 / 1.4 5 0.0 / 7.5 4.7 / 10.7 6 -11.4 / -3.0 0.4 / 5.0 7 -4.2 / 8.8 -12.9 / -7.4 9 -9.9 / -2.2 1.6 / 23.2 LiDAR QA/QC: Experimental Results (II) Dataset Description Photogrammetric & LiDAR Mapping Ayman F. Habib 144 LiDAR QA/QC: Experimental Results (II) Dataset Description Photogrammetric & LiDAR Mapping Ayman F. Habib 145 Strip pairs Flying Direction Rigorous 1&9; 2&4; 5&6; 5&7 Quasi-rigorous 1&9; 2&4; 5&6; 5&7 Simplified 1&9; 2&4; 5&7 Strip pairs Flying Direction % Overlap Average Lateral Distance D (m) Average Flying Height H (m) 1&9 approx. parallel 75 66 130 2&4 approx. opposite 70 160 130 5&6 cross - - 230 5&7 approx. opposite 75 10 230 LiDAR QA/QC: Experimental Results (II) Photogrammetric & LiDAR Mapping Ayman F. Habib 146 1&9 ~Parallel Direction 2&4 ~Opposite Direction 5&6 ~Cross Direction 5&7 ~Opposite Direction Dataset Description LiDAR QA/QC: Experimental Results (II) Photogrammetric & LiDAR Mapping Ayman F. Habib 147 Method δΔω(o) δΔφ(o) δΔκ( o) S/δS Simplified 0.039 0.092 -0.029 -0.00028204 Quasi-rigorous 0.038 0.093 -0.044 -0.00000514 Rigorous -0.094 0.032 90.064 1.00017 Please, note that the estimated parameters are not compatible since different coordinate systems definition are utilized in the two calibration approaches. yb xb zb xlu ylu zlu xb yb zb xlu zlu ylu Flight direction Flight direction Simplified/Quasi-Rigorous Rigorous LiDAR QA/QC: Experimental Results (II) Photogrammetric & LiDAR Mapping Ayman F. Habib 148 Before Calibration After Calibration: Rigorous Approach 1m After Calibration: Quasi-Rigorous Approach 1m Qualitative QC 1m LiDAR QA/QC: Experimental Results (II) Photogrammetric & LiDAR Mapping Ayman F. Habib 149 Before Calibration After Calibration Rigorous Quasi-Rigorous Simplified 1&9 XT’(m) YT’(m) ZT’(m) XT’(m) YT’(m) ZT’(m) XT’(m) YT’(m) ZT’(m) XT’(m) YT’(m) ZT’(m) 0.00 -0.21 -0.07 0.01 -0.01 0.01 0.03 -0.11 0.01 0.04 -0.13 0.00 Ω’(o) Φ’ (o) Κ’(o) Ω’(o) Φ’ (o) Κ’(o) Ω’(o) Φ’ (o) Κ’(o) Ω’(o)

Compartilhar