Multi-target Skeleton Extraction Method of Beef Cattle Based on Improved YOLO v3

被引：0
|
作者：
Zhang H. [1 ,2 ]
Li Y. [1 ]
Zhou L. [1 ]
Wang R. [1 ]
Li S. [1 ,2 ]
Wang H. [2 ,3 ]
机构：
[1] College of Information Engineering, Northwest A&F University, Yangling
[2] Ningxia Intelligent Agricultural Industry Technology Collaborative Innovation Center, Yinchuan
[3] West China Electronic Bussiness Co., Ltd., Yinchuan
来源：
Nongye Jixie Xuebao/Transactions of the Chinese Society for Agricultural Machinery | 2022年 / 53卷 / 03期
关键词：
Beef cattle; Key point detection; Multi-target skeleton extraction; Receptive field block; Target detection; YOLO v3;
D O I：
10.6041/j.issn.1000-1298.2022.03.030
中图分类号：
学科分类号：
摘要：
In view of the problem that the extraction accuracy of beef cattle skeleton was decreased greatly with the increase of targets in the process of beef cattle behavior recognition, an improved YOLO v3 algorithm (Not classify RFB-YOLO v3, NC-YOLO v3) was proposed. After the backbone network, receptive field block (RFB) was introduced to expand the receptive field of the model, and the classification module was eliminated to improve the detection efficiency. Combining 8SH (8-Stacked Hourglass) algorithm to realize multi-target detection and skeleton extraction of beef cattle in actual breeding environment. In the experiment, totally 16 key nodes were set for the beef cattle skeleton to form the beef cattle pose point information, and the detection accuracy was improved through multi-scale and multi-direction training of the image. Based on the statistical analysis of key points of multi-target skeleton extraction model, a method for beef cattle standing and lying down behavior recognition was proposed. Experimental results showed that in terms of target detection, the recall of the NC-YOLO v3 model can reach 99.00%, the precision can reach 97.80%, and the average precision can reach 97.18%. Compared with the original model, average precision was increased by 4.13 percentage points, and the amount of network parameters removed was 13.81 MB; in terms of single-ox skeleton extraction, the 8-Stacked Hourglass network was used to detect key point positions, and the mean average precision can reach 90.75%. In terms of multi cattle skeleton extraction, compared with the model constructed by YOLO v3, the mean average precision of the model constructed by NC-YOLO v3 was increased by 4.11 percentage points to 66.05%. © 2022, Chinese Society of Agricultural Machinery. All right reserved.
引用
收藏
页码：285 / 293
页数：8
相关论文
共 30 条
  • [1] KING M T M, LEBLANC S J, PAJOR E A, Et al., Cow-level associations of lameness, behavior, and milk yield of cows milked in automated systems, Journal of Dairy Science, 100, 6, pp. 4818-4828, (2017)
  • [2] HE Dongjian, MENG Fanchang, ZHAO Kaixuan, Et al., Recognition of calf basic behaviors based on video analysis, Transactions of the Chinese Society for Agricultural Machinery, 47, 9, pp. 294-300, (2016)
  • [3] POURSABERI A, BAHR C, PLUK A, Et al., Real-time automatic lameness detection based on back posture extraction in dairy cattle: shape analysis of cow with image processing techniques, Computers and Electronics in Agriculture, 74, 1, pp. 110-119, (2010)
  • [4] ZHAO Shengbo, WANG Song, LIN Jialin, Clinical symptoms and comprehensive treatment of esophageal obstruction in beef cattle, Modern Animal Husbandry Technology, 8, (2017)
  • [5] SONG Huaibo, LI Tong, JIANG Bo, Et al., Automatic detection of multi-target ruminate cow mouths based on Horn-Schunck optical flow algorithm, Transactions of the CSAE, 34, 10, pp. 163-171, (2018)
  • [6] GU Jingqiu, WANG Zhihai, GAO Ronghua, Et al., Recognition method of cow behavior based on combination of image and activities, Transactions of the Chinese Society for Agricultural Machinery, 48, 6, pp. 145-151, (2017)
  • [7] ZHANG Hongming, WU Jie, LI Yongheng, Et al., Recognition method of feeding behavior of multi-target beef cattle, Transactions of the Chinese Society for Agricultural Machinery, 51, 10, pp. 259-267, (2020)
  • [8] MATTACHINI G, RIVA E, PROVOLO G., The lying and standing activity indices of dairy cows in free-stall housing, Applied Animal Behaviour Science, 129, 1, pp. 18-27, (2011)
  • [9] XUAN Chuanzhong, MA Yanhua, WU Pei, Et al., Behavior classification and recognition for facility breeding sheep based on acoustic signal weighted feature, Transactions of the CSAE, 32, 19, pp. 195-202, (2016)
  • [10] HE Dongjian, LIU Dong, ZHAO Kaixuan, Review of perceiving animal information and behavior in precision livestock farming, Transactions of the Chinese Society for Agricultural Machinery, 47, 5, pp. 231-244, (2016)
123