Research on human retinal cell image restoration

In order to improve the observation effect and further study on the human retinal cell image, image processing technology is used for restoration from the distortion and degradation of images, which are acquired through retinal cell imaging system based on adaptive optics. In this paper, two commonly used image restoration algorithms were introduced: Wiener filter and Lucy-Richardson. On this basis, a semi-blind deconvolution image restoration algorithm based on Sobel operator is proposed. The edges detected by Sobel operator were processed differently from other region, which could both effectively remove the noise, and maintain the image detail well. Finally, frequency domain function and point sharpness function are introduced to evaluate the images which are restored with the above algorithms, the results show that the Sobel based semiblind deconvolution algorithm can achieve quite better results in restoration of cell image.