Feeding detection method and system based on deep learning image processing

Abstract

The invention discloses a feeding detection method and a feeding detection system based on deep learning image processing. The method comprises the steps of acquiring a training set and a test set; constructing a RetinaNet target detection network fusing a CLAHE algorithm and a MobileNet neural network, and training and optimizing the RetinaNet target detection network by using the training set; obtaining a model weight file according to the trained and optimized RetinaNet target detection network, reading the model weight file into an ImageAI library fused with a non-maximum suppression algorithm to obtain a feeding detection model, and verifying the feeding detection model by using a test set until a convergence condition is achieved; acquiring a to-be-detected monitoring video, detecting whether a feeding action exists in each real-time monitoring picture in the to-be-detected monitoring video or not by adopting the feeding detection model, and if the detection result of M continuous real-time monitoring pictures is that the feeding action exists, sending alarm information to the administrator terminal. According to the invention, the feeding situation is reliably and effectively detected, the manpower and material resources for feeding supervision are reduced. The timeliness of feeding supervision is improved. The harm of feeding behaviors to animals and the environment isreduced.

Description

technical field [0001] The application belongs to the technical field of video image processing, and specifically relates to a feeding detection method and system based on deep learning image processing. Background technique [0002] In recent years, with the improvement of people's material living standards, the number of tourists has increased significantly, and the number of zoo visitors has also continued to increase, but uncivilized garden visits have emerged. Articles with headlines such as "Stop feeding them, your enthusiasm may harm them", "Feed becomes garbage and pollutes the environment", "Feeding animals without authorization is not love but selfishness" have attracted widespread attention on the headlines. The zoo has taken many measures to deal with the problem of "feeding is hard to stop". For example, the management team of the Beijing Zoo and the city management jointly patrol and enforce the law, dissuade uncivilized garden behaviors such as feeding animals...

AI Technical Summary

Problems solved by technology

[0003] There are still relatively few researches on camera-based animal feeding detection at home and abroad, and there is no feeding detection method based on deep learning image processing.

Images