Super resolution (SR) is considered as one of the “holy grails ” of optical imaging and image processing,which refers to various methods for improving the resolution of optical imaging system beyond the diffraction limit.Different fr om the registration error and costly problem in multiple subpixel image registration fusion me thod to achieve super resolution,this paper introduces a so-called c ompressed sensing method into super-resolution imaging,exploring the general sparsity of most na tural images,and proposes a single image super resolution method,which can implement super resolu tion from a single low-resolution image to obtain a high-resolution one,without any addi tional information collection.This method can capture sufficient data with a single shot to achieve image super resolution.Simulation results demonstrate that the single image super-res olution method based on compressed sensing proposed in this paper is significantly superior to the co uple d dictionary training method in the sharp details of the image content with lower graininess noise and much less reconstruction time,resulting in both advantages of subjective visual qual ity and objective signal to noise ratio (SNR) metric,which is very simple for realization and has important application prospects,especially for large scale images.