Super-resolution Reconstruction Of Sequence Images Based On Oblique Sampling

With the continuous development of science,higher resolution satellite image is needed.However,limited to the system hardware,it is urgent to propose new system sensing detectors,explore new imaging application modes and methods,and achieve transcendence and development in key technology fields.In recent years,the emergence of sub sampling imaging technology has provided direction for the new mode sensor detector."Sensor based program control technology" came into being.Through the rotation of multiple CCD or CMOS sensor arrays,special innovative sampling that cannot be achieved by traditional imaging methods is achieved.One of the most important steps in imaging based on "sensor primitives program control" is how to reconstruct the sequence image with super resolution according to the special sampling mode of the sensor base element program control technology,which is the key point of the paper.The main methods and contributions of this paper are:(1)For the special sampling mode,a bilinear interpolation algorithm with simple structure and small computation is proposed,which can be implemented on the satellite.(2)For the super-resolution reconstructed image obtained by bilinear interpolation,the Ll fidelity term de-noising model with TV regularization is established,and the alter-nate multiplier algorithm is used to reconstruct the interpolated image.The algorithm is mainly aimed at the edge of the interpolated image.(3)The image undersampling process is regarded as operator A,so the super res-olution reconstruction problem is equivalent to Ax = B,B is the sampled image,and X is the original image.Landweber algorithm is used to solve the inverse problem and improve the resolution of the image.(4)The image reconstruction model with sparse operator is established,and the high resolution image is obtained from the low resolution sequence image by using the alternating multiplier with preprocessing,thus the superresolution reconstruction is realized.(5)For the expensive storage of satellite data,a random sampling mode was pro-posed,an image reconstruction model with sparse operator was established,and the sequence image obtained by CCD or CMOS sampling was reconstructed using an alter-nating multiplier method with preprocessing.In summary,the paper takes the special sampling mode of the sensor primitives program control technology as the background,and carries out the algorithm research in the aspects of super-resolution reconstruction of sequence images and image reconstruc-tion under random sampling mode.Provides new ideas for sequential image processing research...