Automatic Co-Registration of Optical Satellite Images and Airborne Lidar Data Using Relative and Absolute Orientations

摘要

Co-registration of 2-D images and 3-D lidar points in a common area is an important task in data fusion as well as other applications. As the information acquired by image and lidar systems are not the same, the registration of heterogeneous sensors is a challenging issue. The objective of this study is to perform the co-registration of 2-D images and 3-D lidar points using relative and absolute orientations. The proposed method performs image matching between stereo images for relative orientation modeling and generates a matched 3-D surface model. Then, an automatic least squares 3-D surface matching is applied between matched 3-D surface model and lidar 3-D points. Finally, the precise object-to-image transformation and orthoimages can be generated via relative and absolute transformation. The test data include WorldView-2 image, QuickBird image and lidar data. The experimental results indicate that the relative orientation may reach subpixel accuracy while the absolute orientation may reach 1 pixel accuracy. Moreover, the geometric consistency between orthoimages is better than 0.5 m on the ground.