Traditional image enhancement methods often suffer from low image clarity and high distortion.Therefore,a deep learning based enhancement processing method is proposed.After obtaining the monitoring video image information,perform sequence alignment processing on the image.Display the distribution between key pixels and adjacent pixels in a single image through image quantization processing.Using deep learning technology,the directional gradient of the reflection image is extracted using gradient operators in 8 directions to obtain a detail enhanced reflection image,which is then processed through Retinex inverse transformation.Design comparative experiments to demonstrate that this method improves the image enhancement effect of surveillance videos and can obtain clearer scenery images.
关键词
深度学习/监控视频/图像增强/处理方法
Key words
deep learning/surveillance video/image enhancement/processing method
维普
万方数据