Design Of Sheep Number Detection System

Inner Mongolia Autonomous Region is an important animal husbandry production base in China,but the automation level of the meat sheep breeding industry is low.In the breeding process,a large amount of artificial labor is used for management,which not only has low production efficiency,but also affects the overall breeding efficiency,and restricts the rapid development of the industry.In the breeding management of sheep,the herdsmen or farms need to count the number of sheep,which is a time-consuming and laborious problem.A sheep number detection system is designed,which can view the number of flocks in real time.It is convenient for herders and managers to know the number of flocks in time,which has certain practical significance and application value for the intelligent development of the sheep breeding industry.The main work of this paper is as follows:1.The overall architecture design of the system.Design the framework of each part of the system by analyzing system requirements.Set up the software and hardware environment for system development,select the hardware required for the system,such as development board,4G network communication module,camera module for the further development and design.2.Image acquisition system design.Configure the embedded system,including setting up a cross-compilation environment,configuring the Linux kernel,and making the root file system.Design and implemente the functions of image acquisition,image transmission and image receiving module,including the use of key technologies,the preparation of various parts of the program to meet the functional requirements and performance requirements of the design.Finally,the image acquisition system is tested and analyzed.3.Design of sheep detection system based on YOLO algorithm.Using the regression detection algorithm YOLO to extract the position information of the input image sheep by extracting the features of the whole image and using convolution operation and non-maximum value suppression processing.The number of detection frames in the statistically detected image,that is,the total number of sheep.Design user graphical interface to achieve image loading and detection counting.Finally,the test system is tested and analyzed.