Traditional noise removal arithmetic required that the signal-noise-ratio of the image is high, noise removed image by the traditional noise removal arithmetic loss large amounts of edge and texture information in the image. This paper presents a novel procedure for the low signal-noise-ratio (SNR) image noise removal on the local edge-preserving function which makes up the limitation of the traditional methods. Firstly, adaptive median filter is used to remove the part of salt-and-pepper noise and to preserve the edge and texture information in the image. Secondly, local edge-preserving function is built on basic of analyzing the relation of pixels in the local image block. Lastly, a minimization problem is solved by the Poly-Ribière-Polak (PRP) method in order to remove the Gaussian noise and the remnant salt-and-pepper noise in the image. Comparing the results of removing noise, noise removal efficient by our method is better. Especially when the PSNR of the image is 5.4db, the PSNR of the result image is 24.3db.
