Design and realization of an intelligent livestock watering system based on computer vision
[Objective]Aiming at the problems of semi-intelligence,single device function and lack of in-telligent management in monitoring and managing livestock drinking behavior,an intelligent livestock drink-ing water system based on computer vision is designed and implemented.[Method]The Jetson nano em-bedded device is used as the main controller to identify livestock and supply water based on computer vision while video monitoring.The lightweight detection method is used to detect the distribution density of live-stock,and analyze whether it is necessary to add water tanks to avoid congestion.RFID technology is used to identify the tag information and count the key information such as whether livestock individuals drink wa-ter,drinking times and total drinking time.The sensor technology is used to detect the water level and tem-perature,which achieves intelligent control such as power failure when the water tank is full and automatic emptying before freezing.We have developed client and server software,deployed cloud platforms on Al-ibaba Cloud servers,and provided services such as data visualization and tying additional information to live-stock for human-computer interaction.[Result]A 40-day test was conducted on 5 cattle from individual family cattle herds in Inner Mongolia.The results showed that the system can perform video monitoring,object detection,livestock density detection,livestock information statistics and automatic water supply and stop water supply to the tank.Taking the results of the experiment on 8 February 2023 as an example,the five cows tested on the same day all drank water,of which No.2166430 cows had an online time of 3.2 min-utes,drank water once and drank water for 1.3 minutes.According to the drinking water situation of the cattle and the tested population in the past week,the drinking water situation of the tested population is nor-mal.According to the statistics of historical drinking water,the tested group was in a healthy state within 40 days from the perspective of drinking water.[Conclusion]The system reduces the human cost of herding,alleviates the pressure of herding and provides an important basis for herders to judge the health of their live-stock.
computer visionlivestock density detectionintelligent drinking waterlivestock informa-tion statistics
NETL
NSTL
万方数据
维普
