An image denoising method based on BP neural network optimized by improved whale optimization algorithm

Abstract As an important part of smart city construction, traffic image denoising has been studied widely. Image technique can enhance the performance segmentation and recognition model improve accuracy results. However, due to different types noise degree pollution, traditional methods generally have some problems, such as blurred edges details, loss information. This paper presents method based on BP neural network optimized by improved whale optimization algorithm. Firstly, nonlinear convergence factor adaptive weight coefficient are introduced into algorithm ability characteristics standard Then, is used optimize initial threshold value overcome dependence in construction process, shorten training time network. Finally, applied benchmark denoising. The experimental results show that compared with Median filtering, Neighborhood average filtering Wiener proposed better peak signal-to-noise ratio.