Design And Implementation Of Non-Perceptual Weight Monitoring System For Sheep

In the process of livestock breeding,body weight is an important indicator to measure the growth and health of livestock,as well as an important parameter for pasture managers to provide nutrition and make slaughtering schedule.Therefore,long-term weight monitoring of livestock is essential.The traditional method of animal weighing is to tie the animal to make it at rest and then weigh it,or to catch the animal on a scale to make it stable and then read the weight of the animal.This traditional weighing method has many disadvantages,such as serious human interference in the weighing process,which can cause stress to the animal and affect the healthy growth of the animal;as well as long weighing time,low weighing efficiency,wasting a lot of human and material resources,and cannot achieve daily monitoring of the animal’s weight.In recent years,significant research progress has been made in the field of dynamic weighing of animals,which effectively reduces the stimulation of individual animals,but most of them are applied to weighing of individual animals and are not suitable for daily monitoring of sheep weight.Sheep are herd animals,and driving them through the scale one by one will cause them to be frightened and have a strong stress reaction.Moreover,the number of sheep in a farm is large,and passing the scale one by one will waste a lot of time and affect the normal activities of sheep.Therefore,this thesis designs a sheep non-perceptual weight monitoring system,which allows multiple sheep to pass the scale freely at the same time and calculates the weight of each sheep based on the weight data and ear tag information to realize the daily monitoring of sheep weight without intervention.The main research of this thesis is as follows:(1)Through long-term observation of sheep’s daily activity patterns on the farm,the design objectives were proposed according to the needs,and the overall design of the sheep non-perceptual weight monitoring system was developed.(2)The hardware design of STM32-based sheep non-perceptual weight monitoring system.The hardware part of the system mainly includes the design of the weighing platform and the hardware circuit of the microcontroller system,where the weighing platform consists of the mechanical structure of the scale,the pressure sensor and the RFID reader;the hardware circuit of the microcontroller system includes the power supply module,the STM32 core processor module,the serial communication module and the wireless communication module,etc.(3)The software program design of the system.The software part mainly includes data acquisition program and data processing program,the data acquisition program is completed by the microcontroller system,which is mainly responsible for the acquisition of sheep weight and ear tag information;the data processing program is mainly to filter the collected data,and then eliminate the abnormal data based on the random forest algorithm to build a multivariate linear regression equation set,and finally solve the multivariate linear regression equation set based on the average difference coefficient method to get the weight information of each sheep.Through a large number of experiments to verify,the sheep non-perceptual weight monitoring system proposed in this thesis can achieve daily monitoring of sheep weight with certain accuracy and without affecting the normal passage of sheep,which truly achieves unmanned intervention and avoids stimulation of sheep.And the system equipment cost is low,high degree of automation,can adapt to the environmental conditions of the pasture,is an important part of the automatic farming process,has broad application prospects.