Open AccessResearch on the Method of Individual Identification of Chickens Based on Depth Image
Author(s) -
Baoquan Zhang,
Yunlin Qiu,
Xinyu Wang,
Huishan Lu,
Fujie Wang
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1631/1/012018
Subject(s) - artificial intelligence , segmentation , image segmentation , computer vision , pattern recognition (psychology) , scale space segmentation , identification (biology) , image (mathematics) , computer science , image processing , segmentation based object categorization , biology , botany
In order to accurately extract the individual images of chickens from the complex background, the experiment took green shell laying hens which scatter-feed as the segmentation target, and proposed a method of individual identification of chickens based on range image, including image background segmentation and chicken overlapping segmentation. First of all, the depth image after background segmentation is obtained by using fixed threshold segmentation method and morphological processing, and then the image with adhesion is separated by using the method of concave point analysis. Through the recognition of 200 depth images, the results showed that the recognition algorithm can effectively separate the adhesive individuals, thereinto, the correct recognition rate for the image without adhesion is 100%, while the recognition accuracy was 93% for images with adhesion. The comprehensive recognition accuracy was 97%.