Research On Milk Level Measurement Technology Based On Image Processing

China is a big country in milk production and consumption.With the popularization of dairy products nationwide,people pay more attention to the development of dairy industry.The measurement of milk level is of great significance to the development of animal husbandry and aquaculture.At the same time,it also puts forward higher requirements for milk production and measurement technology.This paper designs a new milk level measurement system based on image processing after analyzing and summarizing the common problems in the current milk level detection technology.The implementation goal of the subject is to monitor the milk level in real time and transmit data wirelessly with low power consumption.SKF key frame extraction algorithm is used to extract relevant pictures,and the liquid level image is recognized in real time.The recognition result is output on the PC.Ranch staff can master milking dynamics according to milk level changes,and judge cow mammary gland state through milking duration,thus realizing efficient milking process.According to the date of dairy cows,the individual milk production level of dairy cows is analyzed,and an appropriate feeding scheme is formulated to realize the value of pasture.The foundation of the project implementation is divided into two parts: the hardware part and the software part.The hardware implementation of the project consists of embedded chip(STM32F103),wireless transceiver,CCD camera and AV acquisition card.The software implementation consists of MDK(C language)development environment,Matlab(M language)and Wi-Fi wireless communication technology.The CCD camera is installed at the railing in front of the milk metering bottle.The AV acquisition card is installed on the interface of the development board to obtain the milk liquid level image information.Wi-Fi is used for wireless transmission to a PC end,and the PC end processes liquid level images by calling video information.The PC end uses image enhancement,image segmentation and target extraction in scale area to obtain the liquid level height.Ranch staff can view or monitor through PC.