A Variable Sampling Compressed Sensing Reconstruction Algorithm Based on Texture Information

Aiming at the problems of large storage space, blurred the reconstructed image and poor reconstructed effect under low sampling rate while reconstructing an image with the orthogonal matching pursuit algorithm. A variable sampling compressed sensing reconstruction algorithm based on texture information is proposed. The algorithm divides the image into blocks first, distributes the sampling frequency adaptively according to the texture information and sparsity of the image blocks, and smoothes the reconstructed image using median filtering algorithm to reduce the effect of the reconstructed image blocks. The simulation results show that the subjective and objective reconstruction effects of the algorithm in this paper are better than other reconstructed algorithms under the same conditions, and the PSNR value is $2 \sim 4$ dB higher than other reconstructed algorithms at a low sampling rate, and the higher reconstruction quality is achieved at the low sampling rate.