Rain line removing device in industrial detection

Abstract

The invention provides a rain line removing device in industrial detection, and relates to the technical field of industrial detection. The device comprises a first channel, a second channel and a third channel which are connected in series, wherein the first channel, the second channel and the third channel respectively comprise semantic information extraction modules and feature screening modules, the feature screening modules of the first channel and the second channel comprise channel attention modules and convolution modules, and the feature screening module of the third channel comprises a residual channel attention module and a convolution module. A gradual rain removal structure is adopted, feature extraction is further carried out based on the last extraction result each time, the problem that the rain removal effect of a single stage is poor is solved, the channel attention module is mainly used for further screening voice information in the first two stages, and the residual channel attention module is mainly used in the last stage, residual learning can be carried out on semantic information extracted by channel attention modules in the first two stages, and the feature extraction effect is further improved, so rain line information needing to be removed can be accurately extracted.

Description

technical field [0001] The invention relates to the technical field of industrial detection, in particular to a device for removing rain lines in industrial detection. Background technique [0002] Rain is a relatively common dynamic severe weather. The raindrops fall randomly, the speed of falling is very fast, and they are randomly distributed in the air. In the image, the raindrops mainly appear in the shape of rain lines, which is easy to cause blurred images, loss of detail information, and even Some areas in the image will be randomly occluded, which will affect the performance of the machine vision algorithm. At present, most of the machine vision algorithms used in industrial sites process images. If the images are not derained, the efficiency and accuracy of detection will be seriously affected. [0003] At present, the image deraining method is mainly single-stage deraining method. The single-stage deraining method is to make the rain image in the data set and the...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the above-mentioned technical problems, the first object of the present invention is to propose a device for removing rain lines in industrial detection, which adopts a progressive rain removal structure, and each time further performs feature extraction based on the last extraction result to solve the problem of The single-stage rain removal effect is poor. The first two stages mainly use the channel attention module for further screening of voice information, which can efficiently extract features on the channel. The last stage mainly uses the residual channel attention module, which can target The semantic information extracted by the channel attention module in the first two stages is used for residual learning to further improve the feature extraction effect, so that the rainline information that needs to be removed can be accurately extracted

Images