Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

An in-column mammal counting method based on an instance segmentation algorithm

A mammalian segmentation algorithm technology, applied in the field of computer vision, can solve problems such as complex models, easy missed detection of targets, and poor real-time performance, so as to avoid stress response, fast and accurate counting results, and improve segmentation accuracy

Active Publication Date: 2019-05-21
HARBIN ENG UNIV
View PDF8 Cites 20 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The model used in this method is complex and has poor real-time performance. It only detects the bounding box of the target, which makes it easy to miss the detection of mutually occluded targets.
[0003] In summary, there is currently no simple and practical method for counting mammals that can accurately detect animals in an area and count the number of animals in a breeding factory in real time

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 in-column mammal counting method based on an instance segmentation algorithm
  • An in-column mammal counting method based on an instance segmentation algorithm
  • An in-column mammal counting method based on an instance segmentation algorithm

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0053] The technical solution of the present invention mainly includes the following parts:

[0054] 1) Image acquisition part

[0055] The image acquisition part completes the preservation of mammalian images in each column of the breeding plant. In the present invention, a wide-angle camera is used to collect images. In the wide-angle camera acquisition scheme, the camera is set above the column. By reading the camera parameters, the workstation remotely controls the camera for acquisition work, and the collected images are stored in the workstation.

[0056] 2) Image preprocessing part

[0057] Select an image with a clear target outline, and divide the image into a training set, a validation set, and a test set at a ratio of 8:1:1. Use the annotation tool VGG Image Annotator (via) to annotate the training set and validation set images, draw a closed polygon according to the target outline in the image, and mark the polygon area as a mammalian name, such as "pig". After the image...

Embodiment 2

[0078] figure 1 It is the overall block diagram of the present invention. The experiment conducted by the present invention takes live pigs as an example, starting from the wide-angle camera collecting images of pigs in each stall of the breeding plant, the images collected on site are figure 2 As shown; the annotated image is as Figure 3a As shown; the annotation data of the image is as Figure 3b As shown; the feature map output by the feature extraction network is as Figure 4 As shown; the anchor mapped from the feature map back to the original image is as Figure 5a As shown; the anchor after bounding box regression and non-maximum suppression is as Figure 5b As shown, the dashed box is the original bounding box, and the solid box is the bounding box after regression; the output result of the regional candidate network is as follows Image 6 As shown; the segmentation result of the target is as Figure 7 As shown, the green contour is the labeled contour, and the red conto...

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 in-column mammal counting method based on an instance segmentation algorithm, and belongs to the field of computer vision. The method includes: collecting and sending An image of a mammal in the fence to a workstation through a camera; Selecting images with clear target contours, dividing the images into a training set, a verification set and a test set, and using the training set, the verification set and the test set as data sets for model training; Secondly, generating a segmentation model for testing through an instance segmentation algorithm of deep learning; Sequentially inputting the test set images into a segmentation model generated by training for prediction, and outputting a test result and a test effect picture; After the test is finished, automatically storing the test effect picture; Finally, counting the target bounding box in the test result, thereby achieving the counting of the number of targets in the image. According to the method, the neural network is utilized to train the model, complex image preprocessing steps are avoided, and the identification accuracy of the blocked target is effectively improved.

Description

Technical field [0001] The invention relates to the field of computer vision, in particular to a method for counting mammals in a fence based on an instance segmentation algorithm. Background technique [0002] With the development of industrial integration, large-scale breeding plants continue to increase, and real-time and accurate inventory of mammals in breeding plants has become one of the important issues that large-scale breeding companies need to solve. At present, the number of mammals in the breeding plant is mainly manually counted. The manual count of thousands of mammals is a time-consuming and labor-intensive process, and the accuracy of the count is low, requiring different times of review. The labor cost is very high, much higher than the cost of using equipment statistics, because the equipment is a one-time investment, and the labor cost is a continuous investment. In 2017, the Chinese invention patent (CN206039601U) proposed to count the number of experimental...

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): G06T7/11G06T7/12G06K9/62
Inventor 苍岩陈婵乔玉龙陈春雨何恒翔李诚唐圣权
Owner HARBIN ENG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com