The Algorithm And Software Implementation Of Video 3D Denoise Based On Motion Estimation

Digital video image, which is from the video source to the image video display, must be process by several images processing program. Almost of all procedures will increase the noise and effecting image quality. So the best way to reduce the noise impact is getting rid of the source noise of the image before any other image processing.The difficulty in the image denoise is removing the noise without affecting the original image information. On the other hand, the best denoise is bluring the noisy signal but saving the image contours and texture. According to the previous works, the 2D noise reduction algorithm is difficult to effectively identify between the original image and its noise, because the noise always have strong correlation on image image contours and texture. The noise can’t be removel clearly by only with 2D image information.Based on the above issues, this paper presents a 3D spatial and temporal mixing denoise algorithms, this algorithm contains two algorithm, a 2D frequency domain self-adaptive parameters denoising method and 3D mix-mode noise reduction algorithm. The 2D frequency domain self-adaptive parameter denoise algorithms are denoising by different frequency domain which seperated by different size of low pass filter. With noise statistic modeling method on those separated frequency, the noise can be easily remove. According to the simulation show that the algorithm has a very good ability to remove noise, the picture quality is only slightly worse than the bilateral filter, and the running time was nearly 2.7 times less then the bilateral filter. The algorithm is simple, and suitable for hardware implementation.The 3D mix-mode noise reduction algorithm is to find the most similar MB by use motion compensation technology and analysis the information to set the MB as background or non background one. If the MB is set to background, then average it with current MB with 2D denoised, if not, then process it with 2D filtering method. According to the experiment result, the proposed 3D mix-mode noise reduction algorithm can keep more video image detail and efficient removal of noise outside noise and improve image quality video images visually.