Application of Multi-sensor Data Fusion in Greenhouse

Abstract

With the development of communication, computer and sensor technology, the application of Internet of things technology to agricultural monitoring is the trend of modern agricultural development. Real time and accurate acquisition of farmland environmental information is the basis of precision operation and intelligent management of agriculture, and it is also an important part of agricultural information construction. A farmland environment information monitoring system based on wireless sensor network is designed, crop growth environment parameters are collected by sensor nodes distributed in the field, using CC2530 to build ZigBee data transmission network, the information transmission between ZigBee network, GPRS network is realized by embedded gateway, and the remote monitoring of farmland environmental information is realized. Before data is transmitted, The negligent errors in the measurement data are excluded by Grubbs’ criterion, then the rest of the data are preprocessed based on the arithmetic mean and the batch estimates, lastly the data are fused using adaptive weighted fusion algo-rithm in the condition of minimal mean square error. The results show that the data by hybrid algorithm has perfect accuracy and minimal error. Using this hybrid data processing method, a large number of data can be fused into a data closest to the real situation, and more accurate environmental information can be obtained. The practical results show that, this solution enhances accuracy and reliability of the greenhouse environment detection. This system improves the information level of greenhouse planting, and applys to the management of greenhouse.