Research And Implementation Of Real-time Video Quality Enhancement Algorithm Based On WebRTC

In recent years,with the continuous improvement of Internet bandwidth and the popularity of smart devices,video communication technology has attracted more and more attention from researchers,but the quality of video communication is still unsatisfactory.On the one hand,there is a problem of low brightness in video calls.Currently,most smartphones have video capture functions.However,when the brightness of the room is dark,the brightness of the captured images is poor due to the limitation of the lens aperture of the device.In addition,some mobile phones are lighted by the lens.The lack of brightness causes the brightness of the video image to be dark,which limits the quality of the live video call.Therefore,enhancing the brightness of the video image is an important research direction of this paper.On the other hand,in video communication,people tend to be more interested in skin-related areas in video images,but the current skin color detection algorithm has the problems of poor detection effect and low accuracy.The improved algorithm improves the detection accuracy and performs accurate denoising.Enhancement is also a research focus of this article.After analyzing the advantages and disadvantages of existing skin color detection algorithms and video enhancement algorithms,this paper proposes corresponding improved algorithms.Aiming at the problem that the low illumination condition and the brightness over-light detection algorithm do not work,this paper proposes an improved adaptive tone mapping image enhancement preprocessing method,which enhances the brightness of the portion of the pixel to be tested that is too low.The portion is processed using contrast stretching techniques to return it to the normal detection threshold.Aiming at the problem that the skin color detection algorithm adopts the fixed model,this paper modifies the skin color detection model and adopts the dynamic threshold method.At the same time,the skin color detection threshold setting introduces the brightness component as the independent variable factor,which improves the accuracy of skin color detection.Through experimental analysis,compared with the original algorithm,the proposed algorithm has a significant improvement in the effect.This paper develops a verification system based on WebRTC to implement the improved algorithm proposed above.Aiming at the delay of network transmission that may be brought about by video processing algorithms,a method for dynamically protecting the frame rate of video transmission is proposed.The experimental results show that the proposed algorithm has achieved good results in the WebRTC platform for video enhancement.It has been applied in the program developed by China Mobile.