Classification model of feeding intensity of fish populations based on LRCN

At present, machine vision technology is booming, and has been successfully applied in various fields. Machine vision technology is also gradually applied to aquaculture precision bait casting and other fields. In this paper, we use a long-term convolution cycle network. Compared with the image data set, the video data set adds one-dimensional timing information, and the circular network can effectively process the timing information in the data set. Finally, the long-term convolution cycle network can effectively classify the feeding state of the fish, and the accuracy of the three classification reaches more than 85%.