Research and Implementation of Image Rain Removal Based on Deep Learning

The traditional rain removal algorithm needs to optimize a large number of parameters, and it is only effective for rain of a specific shape, and the model generalization ability is poor. In recent years, the performance of rain removal methods based on deep learning is better than many traditional methods, but there are problems such as incomplete or excessive rain removal, and incomplete texture reconstruction of background details. This paper proposes a rain removal network based on generative confrontation, which connects the high and low frequency parts and integrates them into the model. At the same time, the attention mechanism cyclic neural network is organically combined, which can better preserve the background texture while removing rain. Theoretical can produce better rain streak removal with better color distortion.