| In the process of grain storage, temperature, humidity, moisture, insects and mildew all these will influence the quality and sanitation security of grain storage directly or indirectly. As the opening market of grain exchange, it will accelerate the speed of insects diffusing and infection in grain storage. Not only insects will cause great lost in quantity of storage grain and reduce the nutrition of grain, but also dead bodies and dejection of insects will cause contaminate to storage grain. Life activity of insects will bring more heat which is microbe needed to multiply. Stored grain pest control is a relatively large area, including the prevention of pests in stored grain and governance. The aspect of prevention of pests in stored grain, to be monitor the quantities and types of pests is an essential part.To follow the modern development of grain storage technology, this paper focused on the grain storage pests voice recognition technology which is the new pest detection, the main research work are as follows:Firstly, this article made a soundproof box for acoustical detection of pest in the stored grain with sound insulation and sound-absorbing material,and set up pests sound acquistion system in the soundproof enviroment. Experiments show that the homemade soundproof box sound insulation condition can meet the needs of the experiment.Second, I acquisted-lesser grain borer, rice weevil and Triolium-three pests’sound singals with the sound acquisiton system in the soundproof environment. Collected three kinds of stored grain pests crawling original sound signal frequency domain analysis to obtain the power spectrum, and extracted feature vectors of three pests.Thrid, study the sound which three kinds of pests crawling classification technology. Voice in the extraction of the feature data are classified using BP neural network classifier, and the network optimization design was carried out, using the optimized neural network to classify pests, its accuracy above95%. |
PDF Full Text Request