The Multi-sensor Data Fusion Technology Research In Animal Building Environmental Monitoring System

China is a big country of livestock and poultry breeding industry. The most directinfluencing factor during the process of livestock and poultry’s production is whether afavorable breeding environment can be provided by raisers, except other influencingfactors such as inheritance and nutrition. Since mastering first-hand information oflivestock breeding is the key of controlling the environment, improving air quality, havinga suitable growth condition, ensuring the health and fast growth of livestock andpreventing epidemic disease effectively.At present, most raisers in our county estimate environmental state by using simpleinstruments and turn on and off the environmental control equipment manually accordingto human experience, in which way both inefficient and ineffective. Detection level alwaysrelies on staffs’ actual operation, which result in unsatisfactory livestock house detectionand controlling.Therefore, in order to improve the level of environmental monitoringlarge-scale breeding of livestock and poultry, and promote the further development ofaquaculture, development of a cost-effective, technologically advanced and easy topopularize live, real-time, on-line monitoring of animal building integrated environmentalmonitoring system is necessary.The system will be computer technology, the virtual instrument technology withmulti-sensor data fusion technology combined.The main research subject of this project isintegrated environment of animal housing,based on LabWindows/CVI9.0softwaredeveloping platform, which designed a virtual instrument system software aim atmonitoring internal environment of animal housing. Centered by various environmentalsensors, field monitoring of housing environment can be achieved by this system throughUDPcommunication protocol which sends mutual independent data and information to theupper computer,and achieve a multi-line real-time monitoring sensors for environmentalmonitoring provides the basis for animal building basis and data support.This thesissuggests a kind of monitoring plan with multi-measure points and multi-sensors that can beapplied to the detection of the animal building environment which is multivariablecoupling. This method combines optimal weighting algorithm and D-S(Dempster-Shafer)Theory algorithm to fuse different sensors data in order to make an accurate judgment onthe animal building environment. Thus, it enhances the robustness of environmentalmonitoring and control system and provide a comfortable living environment. Theexperimental result shows that this method improves the reliability and accuracy of theanimal building’s environmental monitoring system.