Training method of neural network for suspended matter salvage control

Abstract

The invention discloses a training method of a neural network for suspended matter salvage control. In the training method, an image which does not contain gas impurities but contains particulate matter impurities is used as a reference image, and the reference image and a training image are respectively input into two convolutional neural networks with the same structure, so that a training feature map and a reference feature map are obtained; the method further includes determining difference values between feature values of all positions in the training feature map and feature values of corresponding positions in the reference feature map, determining all the positions with the difference values smaller than a preset threshold value from the training feature map, and then deleting the feature values of the positions, thus, a feature map capable of focusing on the expression of other image features other than the particulate matter is obtained.

Description

technical field [0001] The present application relates to the technical field of artificial intelligence, and more specifically, to a neural network training method for suspended object salvage control, a deep neural network-based control method for suspended object salvage, and a method for suspended object salvage control. Neural network training system, control system and electronic equipment for suspended solids salvage based on deep neural network. Background technique [0002] A large amount of rubbish usually gathers along the banks of the river, which needs to be disposed of in time to keep the river clean and tidy, which is an important part of environmental protection work. At present, in order to replace the traditional manual salvage method, automatic salvage equipment has been developed for salvaging garbage along the river. These automatic salvage equipment usually use salvage nets from the shore to salvage the suspended matter floating on the river surface, w...

AI Technical Summary

Problems solved by technology

[0003] However, the applicant of the present application found that in the actual use process, because the automatic salvage equipment drives the salvage net to salvage according to the fixed action parameters, it is impossible to control the angle and the actual depth of the salvage net entering the water like manual salvage. The movement of objects under the action of water flow affects the salvage effect

Moreover, because the movement of floating objects in the fluid is very complex, it is basically difficult to adjust the action parameters of automatic salvage equipment through modeling

Images