Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

An Animal Individual Recognition System Based on Video Tracking Technology

A technology of video tracking and recognition system, applied in image analysis, image enhancement, instrument and other directions, can solve the problems of recognition errors and trajectory association errors, and achieve the effect of strong implementability and strong practical application value.

Active Publication Date: 2021-04-06
北京时间煮梦科技有限公司
View PDF12 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Image-based recognition methods are usually combined with visual tracking technology. When individual animals cross frequently in the image, it is easy to have trajectory association errors, resulting in recognition errors

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
  • An Animal Individual Recognition System Based on Video Tracking Technology
  • An Animal Individual Recognition System Based on Video Tracking Technology
  • An Animal Individual Recognition System Based on Video Tracking Technology

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0044] An individual animal identification system based on video tracking technology, which is divided into two parts, an individual animal detection part and an individual animal tracking part, which specifically includes the following steps:

[0045] (1) Animal individual testing part,

[0046] (1.1) Set the frame rate and save the video as an image;

[0047] (1.2) Faster-RCNN model training set production; (1.3) Faster-RCNN model training calls the training network, performs loss calculation, and judges whether the training is convergent;

[0048] (1.4) Repeated iterative calculations, the loss value converges, and a captive animal target detection model based on Faster-RCNN is obtained;

[0049] (2) The real-time tracking part of individual animals,

[0050] (2.1) Input the first frame image into the trained Faster-RCNN-based captive animal target detection model;

[0051] (2.2) each individual animal position of detection model output and the number of quantity initial...

Embodiment 2

[0079] The implementation environment of the present invention is TensFlow1.3.0, CUDA8.0, cuDNN5.1, OpenCV2.4.13. The specific implementation process is divided into two stages: the target detection network training stage and the specific application stage.

[0080] First, the target detection network training phase:

[0081] Step 1: Install the lens at a suitable position above the pen. The lens should be able to capture the overall image of the pen.

[0082] Step 2: Set the collection time interval and collect 30 pictures per second.

[0083] Step 3: Use Labelme software to complete the labeling of animal targets in the collected pictures, and store them in .xml format files to form a training data set for animal target detection in pens.

[0084] Step 4: Retrain the Faster-RCNN detection model using the picture training dataset of housed animals to obtain a Faster-RCNN-based captive animal target detection model.

[0085] Second, the specific application stage:

[0086...

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 an animal individual recognition system based on video tracking technology, which belongs to the field of machine learning. Based on image and video processing technology, the idea of ​​multi-target tracking is applied to the scene of individual animal identification, and individual animal identification is realized by recording the track position coordinates of each animal in the circle in real time. In the specific implementation of the scheme, the Faster-RCNN multi-target detection model in deep learning is combined with the traditional tracking algorithm Kalman filter to solve the difficult problems such as occlusion, trajectory crossing, and poor real-time performance that often occur in multi-target tracking applications. Using the collected massive captive data models, a captive animal detection model based on the Faster-RCNN model is trained. The invention can effectively realize zero contact, no stress, and individual animal identification in the most natural state of the animal individual, and the installation equipment is highly implementable and has very strong practical application value.

Description

technical field [0001] The invention belongs to the field of machine learning, and in particular relates to an animal individual recognition system based on video tracking technology. Background technique [0002] Animal individual identification is the premise and basis for daily management of animals, and is widely used in daily animal feeding management, animal insurance, animal pedigree, file establishment, etc. At present, the methods commonly used to identify individual animals are divided into two categories: physical identification technology and biological identification technology: 1) Physical identification technology, ring, mark, notch method, tattoo method, branding method, dye marking method, subcutaneous burial of microelectronic chips 2) Biotechnology: DNA recognition technology, iris recognition technology, footprint recognition. Foot rings and wing marks are suitable for avian creatures, and their use in mammals is restricted. Notching, tattooing, brandin...

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 Patents(China)
IPC IPC(8): G06T7/277
CPCG06T2207/10016G06T2207/20081G06T2207/20084G06T2207/30241G06T7/277
Inventor 苍岩乔玉龙陈春雨付海玲于德海李志涵陈其航
Owner 北京时间煮梦科技有限公司
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
Eureka Blog
Learn More
PatSnap group products