A Fast Aerial Image Matching Method Using Airborne LiDAR Point Cloud and POS Data

doi:10.11947/j.JGGS.2019.0104

Abstract

Abstract:

A novel tie point matching algorithm of aerial images with the assistance of airborne LiDAR point clouds and POS data is proposed Firstly, the conjugate point searching strategy used in traditional correlation coefficient matching is improved and a fast algorithm is presented. Secondly, an automatic camera boresight misalignment calibration method based on virtual ground control points is proposed, and then the searching range of image matching is adaptively determined and applied to the image matching of the entire surveying area. Test results verified that the fast correlation coefficient matching algorithm proposed in this paper can reduce approximately 25% of the matching time without the loss of matching accuracy. The camera boresight misalignment calibration method can effectively increase the accuracy of exterior orientation elements of images calculated from POS data, and thus can significantly improve the predicted position of conjugate point for tie point matching. Our proposed image matching algorithm can achieve superior matching accuracy with multi-scale, multi-view, and cross-flight aerial images.

Key words: arial image; tie point matching; fast correlation matching; camera boresight misalignment; LiDAR

Yongjun ZHANG,Xiaodong XIONG,Mengqiu WANG,Yihui LU. A Fast Aerial Image Matching Method Using Airborne LiDAR Point Cloud and POS Data[J]. Journal of Geodesy and Geoinformation Science, 2019, 2(1): 26-36.

Figures/Tables 11

Fig.1

Fig.2

Fig.3

Fig.4

Tab.1

Tab.2

Fig.5

Fig.6

Tab.3

Fig.7

Tab.4

References 22

[1]	BALTSAVIAS E P . A Comparison between Photogrammetry and Laser Scanning[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 1999,54(2-3):83-94.
[2]	SHAN J, TOTH C K. Topographic Laser Ranging and Scanning: Principles and Processing[M]. Boca Raton: CRC Press, 2009: 445-477.
[3]	YASTIKLI N, TOTH C, GREJNER-BRZEZINSKA D A . In-situ Camera and Boresight Calibration with LiDAR Data [C]//The Fifth International Symposium on Mobile Mapping Technology(on CD). Padua, 2007.
[4]	IP A, EI-SHEIMY N, MOSTAFA M . Performance Analysis of Integrated Sensor Orientation[J]. Photogrammetric Engineering & Remote Sensing, 2007,73(1):89-97.
[5]	ZHANG Yongjun, XIONG Xiaodong, SHEN Xiang . Automatic Registration of Urban Aerial Imagery with Airborne LiDAR Data[J]. Journal of Remote Sensing, 2012,16(3):579-595.
[6]	SCHENK T . Towards Automatic Aerial Triangulation[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 1997,52(3):110-121.
[7]	KRUPNIK A, SCHENK T . Experiments with Matching in the Object Space for Aerotriangulation[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 1997,52(4):160-168.
[8]	TANG L, BRAUN J, DEBITSCH R . Automatic Aerotriangulation - Concept, Realization and Results[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 1997,52(3):122-131.
[9]	ZEBEDIN L, KLAUS A, GRUBER-GEYMAYER B , et al. Towards 3D Map Generation from Digital Aerial Images[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2006,60(6):413-427.
[10]	WILKINSON B E, DEWITT B A, WATTS A C , et al. A New Approach for Pass-point Generation from Aerial Video Imagery[J]. Photogrammetric Engineering & Remote Sensing, 2009,75(12):1415-1423.
[11]	BARAZZETTI L, SCAIONI M, REMONDINO F . Orientation and 3D Modelling from Markerless Terrestrial Images: Combining Accuracy with Automation[J]. The Photogrammetric Record, 2010,25(132):356-381.
[12]	ZHANG Yongjun, XIONG Jinxin, HAO Lijuan . Photogrammetric Processing of Low-altitude Images Acquired by Unpiloted Aerial Vehicles[J]. Photogrammetric Record, 2011,26(134):190-211.
[13]	LERMA J L, NAVARRO S, CABRELLES M , et al. Automatic Orientation and 3D Modelling from Markerless Rock Art Imagery[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2013,76:64-75.
[14]	XIONG Xinghua, CHEN Ying, QIAN Zengbo . A Fast, Accurate and Robust Image Matching Algorithm[J]. Acta Geodaetica et Cartographica Sinica, 2005,34(1):40-45. DOI: 10.3321/j.issn:1001-1595.2005.01.008.
[15]	ZHAO Xi’an, CHEN Zhixue, LÜ Jingguo , et al. The Scale and Rotating Invariant Auto Stereo Matching[J]. Acta Geodaetica et Cartographica Sinica, 2012,41(1):81-86.
[16]	JI Shunping, YUAN Xiuxiao . Automatic Matching of High Resolution Satellite Images Based on RFM[J]. Acta Geodaetica et Cartographica Sinica, 2010,39(6):592-598.
[17]	ZHANG Guo, CHEN Tan, PAN Hongbo , et al. Patch-based Least Squares Image Matching Based on Rational Polynomial Coefficients Model[J]. Acta Geodaetica et Cartographica Sinica, 2011,40(5):592-597.
[18]	YUAN Xiuxiao, MING Yang . POS-supported Matching Method for Aerial Images between Neighboring Strips[J]. Acta Geodaetica et Cartographica Sinica, 2010,39(2):156-161.
[19]	ZHANG Yongjun, XIONG Jinxin, XIONG Xiaodong , et al. Main Error Source of Vertical Parallax and Compensation Regression Model of POS Data[J]. Acta Geodaetica et Cartographica Sinica, 2011,40(5):604-609.
[20]	CHEN Siying, MA Hongchao, ZHANG Yinchao , et al. Boresight Calibration of Airborne LiDAR System without Ground Control Points[J]. IEEE Geoscience and Remote Sensing Letters, 2012,9(1):85-89.
[21]	WALKER A S . Responses to Users: The Continuing Evolution of Commercial Digital Photogrammetry[J]. Photogrammetric Record, 1999,16(93):469-483.
[22]	XIAO Han, ZHANG Zuxun . Parallel Image Matching Algorithm Based on GPGPU[J]. Acta Geodaetica et Cartographica Sinica, 2010,39(1):46-51.

Test data	Number of image stereos	Method	T₂	T₁	Number of tie points	Number of image points	Number of missing tie points	Number of missing image points	Number of wrongly matched image points	time/s
A	41	Traditional GCC	0.7	—	1257	13571	—	—	—	960.484
				0.01	1257	13571	0	0	0	861.234
				0.1	1257	13571	0	0	0	788.828
				0.2	1257	13570	0	1	0	729.722
		Fast GCC		0.3	1257	13560	0	11	0	680.266
			0.4		1254	13511	3	60	1	647.718
			0.5		1252	13376	5	195	10	626.375
			0.6		1232	12921	25	650	84	609.141
B	41	Traditional GCC	0.7	—	1089	10279	—	—	—	767.563
				0.01	1089	10279	0	0	0	720.703
				0.1	1089	10277	0	2	0	686.562
				0.2	1089	10276	0	3	0	641.859
		Fast GCC		0.3	1088	10271	1	8	0	609.985
			0.4		1087	10240	2	39	1	590.437
			0.5		1085	10120	4	159	10	574.641
			0.6		1070	9725	19	554	87	563.047

Residuals	Accuracy of weighted average horizontal coordinates using points from single strip			Accuracy of weighted average horizontal coordinates using points from all strips
	dX	dY	dXY	dX	dY	dXY
Rmse	1.8693	0.4218	1.9163	0.3945	0.3014	0.4965
Mean	0.3673	0.2995	1.8962	-0.0266	0.2158	0.4490
Maximum	-2.4376	0.8010	2.4877	-0.7450	0.7083	1.0280

Image matching method	Matching results				Time consumed/min
Image matching method	Number of tie points	Number of image points	RMSE of image point residuals (mm )	Proportion of image points with residuals smaller than three times the RMSE		Harris point extraction	Camera boresight misalignment calibration	Image tie points matching	Total time consumed
Traditional method	2367	14780	0.00245	87.6%	14		—	45	59
The proposed method	2201	13511	0.00233	92.0%	14		6	19	39