Increment Information Acquisition In Remote Sensing

Remote sensing,as an important means of earth observation,plays an important role in both military and civilian use.The pursuit of high resolution has not stopped in the development of microwave remote sensing.Spatial resolution is an important parameter of the image's ability to represent the actual scene.In the field of radar remote sensing,it is necessary to realize the change monitoring or target tracking of the scene.In addition to the spatial resolution,the temporal resolution also has high requirements.However,due to the orbit limitation of synthetic aperture radar,it is difficult to achieve high temporal resolution and high spatial resolution at the same time.In view of the above difficulties,this thesis designs and studies the algorithm scheme,and obtains high resolution increment information with a single high spatial resolution image(before scene change)and a series of high temporal resolution low spatial resolution images(after scene change).The study effectively overcomes the shortcomings that high spatial resolution image always has low temporal resolution.The thesis is mainly consisted of the following:First,the observational degradation model of radar at different resolutions is proposed.Firstly,from the perspective of image degradation,the various factors affecting the imaging quality are analyzed for each part of the radar imaging process.Then according to the proposed model,the specific scheme of this paper is designed,and a series of parameters in the algorithm can be adjusted and controlled.So that the applicability of the algorithm is greatly improved.Second,an incremental information acquisition method based on sparse theory is proposed.This paper finds the inherent sparse characteristics in the scene change situation,and then combines the acquisition process of low-resolution images with the compressed sensing process in sparse theory.And the thesis designs mask for the observation process to achieve effective extraction of incremental information.Third,based on the motion characteristics,the incremental information based on the sparse theory is further modified.For some situation,where the incremental information does not satisfy the strict sparsity,the objects' motion information is excavated to further modify the reconstructed incremental information.Then project image by combining the observation model before and after the change to improve the reconstruct accuracy.Fourth,an effective image reconstruction quality evaluation system is designed,and the experimental performance of the proposed algorithm is verified by experiments.In the evaluation of the information extraction effect,this thesis uses both subjective and objective method to score the reconstructed scene,so that the algorithm results can meet the strict requirements of the human eye and also have good objective performance.Finally,according to the evaluation system design experiment,the performance of the proposed algorithm is analyzed in detail.