Abstract: Image acquisition process,affected by the imaging system,cannot get all the information of the original scene．Super-resolution image reconstruction is the technique to boost image quality under the premise of not changing the imaging system．POCS (projection onto convex sets algorithm) can take advantage of multi-frame low-resolution ima-ge to reconstruct a high resolution image．However,the traditional POCS algorithm usually produces the “jaggy” edge．In nature images,there will be many similar-edge structure．Taking advantage of local similarity structure can effectively eliminate the “jaggy” edge．Therefor a non-local POCS-based super-resolution image reconstruction algorithm was proposed．It can sharpen image edge effectively,and improve the visual senses of the image.

Key words: Super-resolution reconstruction,POCS (convex set projection),Non-local mean filtering,Image interpolation