Image processing method and device, computer readable storage medium and terminal

Abstract

The invention discloses an image processing method and device, a computer readable storage medium and a terminal. The image processing method comprises the following steps: acquiring a to-be-processed image; inputting the to-be-processed image to a trained neural network model, wherein the neural network model comprises a plurality of encoders and a plurality of decoders, different encoders output feature maps with different scales, each encoder fuses feature maps with different receptive fields or overlaps different color channels, and each decoder is used for recovering the encoded feature map and outputting a processed image; and outputting the processed image. According to the technical scheme, the image denoising and brightness improving effects can be improved.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to an image processing method and device, a computer-readable storage medium, and a terminal. Background technique [0002] High-quality images guarantee the accuracy of image post-processing and transmission information. In low-light environments, high-quality and high-contrast images cannot be directly obtained through equipment. In actual application, the current low light enhancement method (Low Light Enhancement, LLE) faces the situation that low-light images have cold tone and noise, and cannot restore details. The former will affect the visual viewing experience, and the latter will lead to information transmission. mistake. [0003] Most of the current low-light enhancement methods are based on RGB image processing, including histogram equalization, tone curve mapping, Retinex method, wavelet transform, and deep learning methods based on rain and fog remova...

AI Technical Summary

Problems solved by technology

[0004] However, the methods in the prior art can only enhance the RGB image, and the details captured by the original sensor (sensor) have been lost in the process of obtaining the RGB image, and the details cannot be restored through the low-light RGB image.

Moreover, the network structure for extracting features has a small receptive field and limited network capacity, which is not effective for image information extraction.

Images