Three-dimensional positioning using ALOS/Prism triple linear-array satellite images

Three-line scanner satellites obtain in-track triple images using multiple linear charged-couple devices (CCDs). These highly overlapped triple images provide valuable intersection geometry and similar features between images and are therefore beneficial to three-dimensional (3-D) modeling. This study focused on orientation modeling and space intersection for in-track triple images and compared the accuracy of independent and block-adjustments using different models. The independent adjustment compensates the orbital error of each image using time-dependent polynomials equations and compensates image-space error with parameter transformation. Block adjustment corrects the orbital error of triple images simultaneously. The test data are ALOS/PRISM triplet images. The 3-D positioning accuracies of the independent adjustment model achieved 6.7 m and 6.5 m in horizontal and vertical directions, respectively. By contrast, the block adjustment achieved 7.7 m and 4.6 m accuracies in horizontal and vertical directions, respectively. Experimental results showed that the accuracy of positioning between independent and block adjustment is similar, but only 10 control points were used in the block adjustment. The proposed block adjustment significantly reduced the demand of the control points.