Application of Multi-Scale Convolution Neural Network Optimization Image Defogging Algorithm in Image Processing

To improve the ability to detect and identify smog images in complex road traffic scenes, need be defogged, an optimized image defogging algorithm on basis of multi-scale convolutional neural network (MCNN) is proposed. The physical model scene scattering constructed, divided sky area, surface area boundary area. line between dark channel established by Canny edge detection MCNN optimization, subjected detail compensation gray enhancement processing through prior knowledge. After substituting atmospheric light value transmittance map into (ACM), learning combined realize filtering optimization scenes. color saturation, degree, peak signal-to-noise ratio (PSNR), texture effect as well other aspects are taken test indexes for simulation experiment. results show that degree definition defogged haze scenes higher using this method, which improves output PSNR has a good application defogging.