Resnet-3D convolution cattle video target detection method based on balance loss

Abstract

The invention discloses a Resnet-3D convolution cattle video target detection method based on balance loss. The method comprises the following steps: 1, segmenting an input original cattle video sequence into frames to obtain a frame picture data set, and labeling the frame picture data set; dividing the marked frame picture data set to obtain a training set and a test set; 2, selecting a plurality of continuous frame pictures in a sliding window manner, and sequentially obtaining a continuous frame picture sequence, thereby expanding and enhancing the data set; 3, performing classification regression and position regression on the obtained frame picture sequence through a target detection network Faster rcnn to obtain a target detection model; and 4, inputting the test video into the trained target detection model to obtain a cow detection frame and confidence thereof. According to the invention, the problem that high-density cattle cannot be detected due to shielding can be effectively solved. Meanwhile, the video is used as a test, and the video is output through the network model, so that the problem that the target is temporarily lost when the image is fused with the video can be effectively solved.

Description

technical field [0001] The invention belongs to the field of computer vision, in particular relates to target detection, and specifically provides a Resnet-3D convolution cow video target detection method based on balance loss. Background technique [0002] Animal husbandry is a traditional industry in our country. In recent years, with the improvement of computer level and the rapid development of computer vision technology, the application of object detection technology in livestock breeding has received more and more attention. However, due to the high density of cattle in livestock farms and the harsh environment of pastures, it is difficult to detect cattle in a natural environment. The target detection algorithm based on deep learning can better extract the characteristics of the cow target, and the detection effect will be better. Applying computer vision deep learning algorithms to cattle detection will help promote the development of large-scale farming in China, g...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to solve the problem of missed detection caused by occlusion between high-density cattle herds and the problem of short-term loss of targets caused by image fusion video in image target detection

Images