Single Image Rain Removal Method Based on Convolutional Neural Network

Abstract

The invention provides a single image rain removal method based on a convolutional neural network and relates to image processing. The method includes the following steps that firstly, manual rain addition is conducted on clean and clear images through a screen blend model to form corresponding rain images, and an image library is built; secondly, the system structure of the convolutional neural network is determined; thirdly, corresponding rain image blocks and rain-free image blocks with the size being 64*64 are obtained from the first step and serve as training samples for training; fourthly, the single image blocks are obtained in an overlapped mode and input in a trained rain removal filter system to obtain the corresponding rain-free image blocks, and weighted averaging is conducted on the image blocks to obtain rain-free images. The method solves the problem that a single image rain removal method based on dictionary learning is long in consumed time, achieves rain removal and guarantees definition of background images, the rain-free images can be quickly obtained after the rain images are input, and the requirement of embedded equipment for real-time processing is met.

Description

technical field [0001] The invention relates to image processing, in particular to a method for removing rain from a single image based on a convolutional neural network. Background technique [0002] The existing single image deraining methods are mainly divided into the method based on dictionary learning and the method of filtering by using filters such as guided filtering. [0003] The method based on dictionary learning considers that the rain line and the background edge belong to different structures, and they should be represented by different dictionaries, so as to distinguish whether the edge belongs to the rain line for deraining. But in real life, some properties such as the direction color of the rain line and some background edges sometimes overlap. In the dictionary separation step of this type of method, although new features are continuously introduced to increase the discrimination of dictionary classification, the accuracy rate is improved to a certain ex...

AI Technical Summary

Problems solved by technology

But for the task of directly removing rain, traditional filters cannot achieve the desired effect

because they can only take into account smaller local neighborhood information

In a small local window, the structural distinction between rainlines and edges is not high enough, and they are more likely to be similar, so they cannot be distinguished well. The processing result is often a compromise between clarity and rain removal, and it is impossible to remove rainlines. While preserving the information of the background image well

Images