Research And Development Of Intelligent Diagnosis System For Common Diseases Of Rice And Wheat Leaves

As the most important food crops in our country,rice and wheat are closely related to people’s lives.Rice and wheat diseases are the biggest stumbling block to the healthy growth of rice and wheat.To accurately prevent and control diseases,accurate diagnosis is an indispensable and important prerequisite.Traditional rice and wheat disease investigations are done by field observation and manual statistics,and there are many unobservable problems.For example,the accuracy of artificial naked eye observation of diseases is not uniform,the classification of diseases is inaccurate,and the disease statistics lack timeliness,which seriously delays the optimal control time.As a result,the output of rice and wheat is severely reduced,causing huge economic losses.Empirical use of pesticides has caused a large number of pesticides to be used,destroying the ecological environment,and may also lead to varying degrees of food safety problems.The development of machine vision and computer technology provides a practical way for machines to replace human observation.This dissertation focuses on the shortcomings of traditional manual investigations.Based on image processing methods,the diseased spot images of rice and wheat leaves are segmented and feature extracted,and the BP spatial segmentation algorithm of disease images and effective multi-scale local sub-feature extraction methods are studied.In order to improve the accuracy of disease classification and recognition,the thesis conducted research based on deep learning methods,and proposed an efficient disease recognition algorithm based on transfer learning,and the effectiveness of the algorithm was verified through experiments.On the basis of the research algorithm,the software and hardware for the automatic diagnosis of rice and wheat diseases were designed using the key technology of the research,and the intelligent diagnosis system for common diseases of rice and wheat leaves was developed.Using machine precision measurement instead of manual investigation,scanning rice or wheat leaves,or using an external camera to take pictures,and then start rice disease diagnosis or wheat disease diagnosis,and get rice and wheat disease type,disease grade data and corresponding pictures.Quick and accurate diagnosis of rice and wheat diseases,and also has the function of uploading the diagnosis results to related platforms.A large number of actual test results on the spot show that the intelligent disease diagnosis system developed has a higher accuracy in the diagnosis of five kinds of diseases of rice and wheat,and its efficiency is better than that of manual labor.The intelligent diagnosis system developed in this paper can effectively replace manual labor,and provides strong support for standardizing the disease monitoring process,precise application of pesticides in farmland,and intelligent and precise diagnosis of rice and wheat diseases.