IDENTIFICATION METHOD OF RICE LEAF BLAST USING MULTILAYER PERCEPTION NEURAL NETWORK
Rice is one of the major cash crops in the world.Rice blast is considered to be the most important disease of rice plants worldwide including China.Among different types of blast, leaf blast often causes the death of young plants during its growth stage.However,excessive use of pesticides for rice disease treatement increases costs and raises the danger of toxic residue on rice plant.Therefore,proper disease control measures must be undertaken in rice growing region to minimize losses.Now,using digital image processing and analysis method are being investigated to early detection of leaf blast.In this study, three multilayer perceptron (MLP) classifiers were designed for classifying the healthy and diseased parts of rice leaves, and we obtain three classification results by using three different input unit, at first, the color and texture features are used as input parameters separately, and then,both the color and texture features are combined as input parameters.Through comparing the results of three MLP classifiers,conclusion can be drawed that if using the both combined color and texture features as input unit, the identification ratio is higher than using color or texture features respectively.The results lay the foundation for automatic diagnosis of rice plant diseases.
leaf blast color feature tezture feature multilayer perception diease identification
Liu Libo Zhou Guomin
Institute of Agricultural Information,Chinese Academy of Agriculture Sciences,China College of Mathm College of Mathmatics and computer,Ningxia University,China
国际会议
北京
英文
1-7
2009-10-14（万方平台首次上网日期，不代表论文的发表时间）