Using Compressive Sensing Technique in Image Sensing

This article presents a new structure of image sensor that compresses the image information based on the compressive sensing (CS) technique and employing a measurement matrix. In the proposed structure to perform compression, the image related voltages on the capacitors in each neighboring pixels are summed in a specific regular pattern. Also, the high sparsity of moving objects images which obtained from the difference images has been exploited and the proposed structure compresses the image information both in the normal and difference modes. Implementation of the measurement matrix and linear combination of the pixel data is done by simple connections between the pixels. The results of 300 images and 30 difference images compressed and recovered through a deterministic measurement matrix show the PSNR values of 26.65 dB and 40.09dB respectively. The results show the effect of applying presented method on the quality of recovered images.