Analysis And Processing Of The Camera Image Noise

Actually, the video images produced by the camera will be affected by various factors, including the noise pollution, fuzzy, distortion and so on, the noise is the most common and main reason of influence the image quality. Generally, the camera can be divided into CCD camera and CMOS camera. Because of different camera response functions, different camera will produce different noise in the process of formation images. To accurate estimation the image noise can provide reliable basis for image denoising.In the phase of noise estimation, this paper uses a model to simulate the camera image noise first, and then uses the camera response function, provided by the computer vision lab of Columbia University, to get a noise level function library. PCA analysis has been made on the function library to get a dictionary. Secondly, this paper uses structure analyzer to search the smooth piece in the noise image, and then uses the smooth piece to estimate the noise sample point of the image. Finally, this paper uses the dictionary and noise samples to estimate the noise level function on the basis of sparse representation.In the phase of image denoising, this paper introduces the block matching3D filtering algorithm and non-local means filtering algorithm, which have good effect of denoising so far. For monitoring video, this paper puts forward a new denoising algorithm based on the Guide filtering and non-local means filtering. And then this paper puts forward the adaptive denoising method based on noise detection. Finally, this paper uses the two methods in the actual monitoring video, and proves the effectiveness through different experiment results.