Research On The Key Theory And Technology Of Effects Of Topographic Relief On The Push-broom Remote Sensing Stereo Imaging

The push-broom remote sensing stereo imaging is the widely used remote sensingimaging type now, which is generally realised through the CCD camera with thelinear array push-broom way. Because of the special imaging mechanism, thepush-broom imaging precision is affected by many factors, such as the topographicrelief, the attitude variation of CCD cameras and the earth rotaton. The topographicrelief will introduce space information loss and imaging error among different viewimages of the push-broom remote sensing stereo imaging. However, the report ofresearch related to topographic relief is rarely seen. Key technology and theory of theeffect of topographic relief on the push-broom remote sensing stereo imaging wasstudied in this paper. The main research are listed as follows:1. The following subjects were systematically and deeply summarized in thispaper: the push-broom remote sensing stereo imaging technology and its currentdevelopment situation and research hot spot, the rigorous sensor imaging model andthe generalized imaging model, the application scope and precision of each imagingmodel.2. The mathematic model between the topographic relief and image disortion ofthe push-broom remote sensing stereo imaging was studied and built. The Gaussiansynthetic surface model was used to simulate the terrain topography of remote sinsingimaging areas. Based on the Gaussian synthetic surface model, the geometricdistortion of the push-broom remote sensing stereo imaging caused by topographicrelief was studied and simulated. The Chang’E-1satellite images were used to provethe analysis results.3. The geodetic coordinate variation of ground points was introduced to expressthe topographic relief so as to solve the problem of the uncertainty about themathematical expression model of the real terrain area. Under the assumption that thethe attitude and orbit of the satellitebone CCD is stable, linear processing of therigorous sensor imaging model was carried out to build the error model between thepush-broom remote sensing stereo imaging and the topographic relief by using the total differentiation method. The push-broom remote sensing stereo imaging errorcaused by different topographic relief was calculated by using the Chang’E-1satelliteimaging parameters and the calculation results were analisized in this paper.4. The modulation model between ground resolution of different viewingimages of push-broom remote sensing stereo imaging and topographic relief wasstudied and built. The global weighted POCS (Projection onto Convex Sets) superresolution reconstruction algorithm based on the image information entropy weightwas proposed. The weighted residual constraint convex set and the weightedrelaxation projection operator were designed by normalizing the information entropyof the original observed remote sensing images. The motion parameters of adjacentframes were estimated by using the2-D affine transformation method. The GaussianPSF (Point Spread Function) was used to revise the reference frame. Experimentresults based on the Chang’E-1images show that, the given algorithm in this paperhas the following advantages:(a) compensating the space information loss of remotesensing images with different viewing angles caused by topographic relief;(2) beinggood at keeping the image texture details;(3) better convergence speed and stability.5. The high-precision standard terrain entity model was bulit based on realremote sensing datum. A push-broom remote sensing stereo imaging simulationmeasurement method based on semi-physical simulation was proposed. The CCDorbit variation was simulated by the translation axis with three degrees of freedom.The CCD attitude variation was simulated by the rotation axis with three degrees offreedom. The method can acquire the CCD attitude and orbit data in real-time withparallel change or distributed change mode. The push-broom remote sensing stereoimaging simulation test on the high-precison terrain entity model will be realisedthrough the given method in this paper.