Design And Implementation Of Body Temperature Detection System For Broiler Breeders

White feather broiler is an important chicken supplier in China.As the parent or grandparent of white feather broiler,the white feather broiler breeder has high economic value and production value.Body temperature is a key physiological index of disease for livestock and poultry.It is helpful to improve the welfare of the flock and the level of precision breeding by monitoring the body temperature of white feather broiler breeder,which can improve the production efficiency.However,the traditional artificial measurement or contact temperature measurement can cause stress or damage to the broiler,and big measurement errors.In this paper,white feather broiler breeder was taken as the research object which has high economic value.Infrared thermal imaging technology,sensor technology and deep learning technology were used to establish a temperature inversion model of the temperature under wings.The temperature of chickens was measured by non-contact method.Finally,a monitoring system of white feather broiler breeder body temperature based on Android APP as a visualization platform was developed.The main research contents were as follows:(1)Broiler house environmental factor sensing system was constructed:MSP430F149 chip and environmental factor sensor module were used to develop nodes and gateways.A wireless sensor network(Wireless Sensor Network,WSN)was made by nodes and gateways through Zigbee.The gateway uploaded data to the cloud server through the communication technology in narrow band Internet of Things(NB-IoT).(2)The inversion model of white feather broiler breeder body temperature under wings:The temperature data of broiler’s regions of interest(ROI)from thermal infrared data was collected by training the YOLOv3 recognition model.The temperature of under wings was collected by placing wireless thermometers.The environment temperature and humidity,light intensity and ROI temperature were used as input,and the temperature under wings was used as the target.The model was trained by multiple linear regression and BP neural network respectively to establish the broiler body temperature inversion model.The performance of the two models were evaluated and compared.The model based on BP neural network was selected as the inversion model of temperature under the wings of white feather broiler breeder.(3)The temperature monitoring software platform for white feather broiler breeder was developed:The middleware was design based on the Java,which was responsible for receiving,parsing,and storing the data of environment and broiler temperature.The database management system was designed based on MySQL for the broiler data of temperature and environmental parameters.An APP for data visualization was developed which was based on the Android platform and the Java.This APP used the MVP architecture to provide users with a well-functioning,concise and practical operation interface which could display information such as broiler house,v body temperature,and environmental factor.(4)The body temperature changes of white feather broiler breeder was researched:In order to test the reliability and stability of the system,the body temperature detection system of white feather broiler breeders was used to research the regularity of body temperature change during the daily behavior of white feather broiler breeders.The common behaviors of white feather broiler breeders include laying,mating and attacking.The rule of temperature change of white feather broiler breeder during their common behavior was analyzed and summarized by drawing the temperature curve of behavior time which could judge the behavior and health status of broiler.