Super resolution reconstruction of three view remote sensing images based on global weighted POCS algorithm

The spatial information compression and data unbalance of CE-1 three view images, which caused by the three-line-array CCD stereo camera imaging view angle and the gradient variety of the lunar surface landform, were seriously considered and an improved super resolution reconstruction algorithm based on global weighted POCS for remote sensing images was proposed in this paper. The global weighted coefficients were calculated by normalizing the mean gradient of original remote sensing images. The residual error and gray scale were introduced as constraint sets during the iteration reconstruction of the three view remote sensing images, by which, the spatial information of remote sensing images was mined and the reconstruction image resolution was compensated and enhanced. Experiment results showed that, the reconstructed images by using the global weighted POCS algorithm have better subjective discerning effect of image details and higher spatial resolution, compared with the widely used algorithm, such as the nearest neighbor interpolation POCS algorithm.