Detection Methods Of Stored-grain Pests Based On Image Processing

China is the largest country of grain production,storage and consumption in the world. Doing well stored-grain management is a very critical thing concerned with the national economy and the people's livelihood. In recent years,the stored grain is more than 500 billions kilograms in our country.The centeral government offers billions of RMB subsidies to the grain storage. But abundant stored-grain is still losing because of the ill management and loss ratio is about 0.2% in the national granary.The stored-grain pest is one of the determinant factor."The Fifth Grain Trade Science and Technology Development Programming"put forward definitely to realize automatic detection about stored-grain pests. The approaches of the Sampling,the Sound Detecting,the Near Infrared and others existed approaches can't provide the grain pests'categories,densities and other parameters.In addition, with the increase of the stored-grain pests'drug fastness,their categories and densities are increasing in recent years.As a result,developing a kind of scientific,precision,simple detection technology for stored-grain pests is very necessary and imperious.There are a series of advantages making use of the image recognition technology to detect the stored-grain pests,such as high precise,low price, high efficiency,no pollution,less labor,convenient connection with the computer grain detection in granary,and so on.It can help the administrators to make scientific decisions so that they can take rational prevention-measures in time,the storage can be managed in quality,quantity and greenness.Stored-grain image and it's pre-procession are the basement for the following work.The machine vision system uses the CCD camera to get the images of stored-grain pests.During the transmission some noises affect the images and the images are abundant in data,In the paper some methods based on singular value decomposation and general matrix inverse are proposed to compress and restore the stored-grain images.Then enhance and smoothen the images by wavelet transform,noise filter and so on.Then each the simulation results are analysied and compared.In order to detect stored-grain pests,some algorithms are applied to remove the grain background by fixed,experience,iteration,biggest variance,automatical threshold gain,fuzzy cluster and simulated annealing.Then the geometric,invariance moment and texture features of stored-grain insect pests are extracted.As a result of feature extraction and selection,the dimension of feature is reduce to 3.Finally,the nearest neighborhood classifier and enclid distance with weights classifier are designed to recognized main stored-grain pests and the correct recongnition is about 95%.