Aerial And Space Line Sensor Positioning Accuracy Analysis And Control Of Image Geometric Quality

The rapid development of world economy leads to the expansion and changing of urban areas, which promotes the increasing requirements for land cover monitoring and change detection and thus brings more chances for photogrammetry applications. However, continuously emerged new platforms and sensors bring higher requirements for the photogrammetric data processing. This paper researches the positioning accuracy and image geometric quality control of airborne and space-borne linear array sensors from three aspects, i.e. platform, mathematical model and terrain.Compared with frame images, the quality of linear array images is easier to be influenced by the unstable platform motion. The jitter of platform will cause the discontinuity between adjacent scan lines, which is the main difficulty in linear array image processing. Based on the rigorous imaging model, influencing factors of imaging geometry quality are analyzed for linear array image, including platform speed, attitude change and so on. Then, formulas are given to express the relationship, according to which, advices of image acquisition control are provided, so as to improve the quality of images for subsequent work.Geometric quality control program for satellite linear array image is designed and realized based on simulation method. This paper intends to provide basis for the optimal satellite designing parameters and program feasibility demonstration, according to the principle of program feasibility demonstration and the characteristics of linear array image. This work is done by conducting simplified simulation and geometric accuracy analysis of satellite photogrammetric system based on the corresponding technical indicators. Then satellite system simulation process is designed and realized, in which laser distance observation data is taken into the adjustment models as parameter according to the need of the space-borne large-scale (1:5000) survey camera program pre-research. And technology system combining single line scan camera and Multi-beam laser ranging is verified. It is proved both theoretically and practically that the use of laser data makes adjustment more stable and reliable, and is conducive to reducing the number of necessary control points and decreasing production costs. The technology system combining single line scan camera and Multi-beam laser ranging fulfills the requirements of1:5000scale surveying and mapping, which means that the project is feasible.To deal with ill-posed problem of RFM (Rational Function Model) parameters estimation caused by over-parameterization, a novel method for automatic RFM parameters optimization by stepwise selection of the necessary parameters based on scatter matrix and elimination transformation has been proposed. The proposed method can fit the rigorous sensor model of HRS imagery with least essential parameters, therefore the ill-posed problem can be significantly alleviated and the rational function model is optimized. SPOT5satellite data is used in the experiment of this part. Compared with the ridge estimation method, the proposed optimization method only makes use of35selected parameters and improves the accuracy by10%-20%. Furthermore, condition index of the normal matrix, which is used for evaluation of rational function model parameters, is reduced from E+7to less than about2000. In a word, the experiment indicates that the over parameterization problem is significantly alleviated.Considering that linear array imaging can be greatly influenced by terrain elevation and slope, this paper gives a proposal that pre-design the platform height and scanning angle according to the terrain, and provides formulations for this purpose. One characteristic of push-broom imagery is the perspective projection in sample direction and the quasi-parallel projection in line direction. Thus, in different terrain, linear array imaging would suffer over-sampling (when ground below average plane) and under-sampling (when ground over average plane) in platform motion direction. This paper studies on the influence of terrain elevation and slope on the linear array imaging both theoretically and actually, proposes that terrain platform height should be designed according to the terrain elevation. The formulations for this purpose are given. The proposal is based on the analysis that if focus length, scanning angle and CCD size are fixed, elevation of n times the Z (ground average height) can cause image point error of n pixels. Besides, an equation is given to describe the relationship among the slope, the scan direction and ground sample distance (GSD). According to that, the scan direction can be well designed to optimize the imaging geometric quality.