Improvement Of Image Denoising Processing Of Mobile Communication Equipment Based On Android

Camera module is the most thick part of mobile communication device,its volume is shrinking,in the current period of time.Area of aperture and photo-sensitive element in camera module are affected by this trend as well,this area is difficult to expand,which lead to the weak sensitivity,and lead to numbers of dark spots in the image acquired in low illumination environment.Normal-ly,the image intensity is weak.Using this feature we propose the concept of intensity factor.And we use intensity factor to improve the existing algorithm,so that the algorithm can enhance the intensity in dark region of the image,to smooth dark spots.In addition,prolong exposure time can accumulate charge.Using this theory,we get sample images for a period of time,to get one im-age.In this way,the denoise ability in the weak illumination environment can be further promoted.Experimental data show that the method proposed in this thesis is effective.The main contents of this thesis are as follows:(1)Applying the feature that images acquired in low illumination has a low intensity,we get an improved curvature model based on a nonlinear diffusion model driven by curvature and edge stopping coefficient.This model utilizes the curvature characteristics of isophote and the edge information,repair the noise and preserve edge.Utilizing the intensity factor,proposed in this thesis,this model could make different scale smooth for different light intensity area.Then ameliorate the dark spot phenomenon caused by insufficient illumination.(2)We use SIFT to detect and locate the feature points for sample images.Calculate the main direction and the auxiliary direction of the feature points.Generated feature vector,and filter them.Using the European distance to match the feature vector between the main image and the auxiliary image,then get the matching point pairs.(3)Using RANSAC algorithm to filter the matching point pairs,using them to get the homography matrix.Then using the homography matrix to do the perspective transformation for main image.Calculate the corresponding points from all of the auxiliary image for a main image pixel point,and take the maxi-mum value of them as the pixel value of the target image.Eliminating the noises using the improved curvature model,after calculate all pixel’s value.And an image with good visual effects is obtained.