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

A target detection and convolution technology, applied in the field of computer vision, can solve the problems of short-term loss of targets and missed detection, and achieve the effect of solving the short-term loss of targets

Pending Publication Date: 2021-04-06
HANGZHOU DIANZI UNIV
View PDF0 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

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

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Resnet-3D convolution cattle video target detection method based on balance loss
  • Resnet-3D convolution cattle video target detection method based on balance loss
  • Resnet-3D convolution cattle video target detection method based on balance loss

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0038] The specific embodiments of the present invention will be further described below in conjunction with the accompanying drawings.

[0039] The invention can acquire the timing relationship between video frames and improve the ability to detect high-density cattle herds. Since some cattle have fewer back patterns, it is difficult to extract feature information. Resnet with a high number of layers can extract very fine detail features, which can greatly improve the ability to extract cattle features. 3D convolution can link the relationship of video context, not only can solve the problem of cattle occlusion in high-density situations, but also solve the problem of short-term loss of targets on consecutive frames when image fusion video is performed to a certain extent. And in order to solve the sample blur phenomenon that may occur in the video frame extraction process, Balanced L1loss is introduced as the supervision function of the regression frame, so that the blurred ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

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...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/00G06K9/32G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06V20/40G06V20/46G06V10/25G06N3/045G06F18/2415
Inventor 李琦沈雷何晶
Owner HANGZHOU DIANZI UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap