Research On Cucumber Disease Identification Based On Image Processing And Pattern Recognition Technology

Vegetable as important economy crop of our country occupies quite a few proportion inagricultural production. Pest is the main limiting factor in vegetable production, and misusingpesticide lead to vegetable high pesticide residues and environment pollution, so identifyingvegetable disease accurately and preventing effectively is the key technology.This article first has introduced the actuality of crop disease identify and pointed out it'snecessity, then taking computer image processing as main technique measure, has studied themethods of automatic recognition in vegetable disease image, integrated image processing,fuzzy mathematics and genetic algorithmTaking the most common cucumber disease as the example, and main research work is asfollows:1. Collected 168 disease specimen using digital camera depend on the advice of experts,including downy mildew, powdery mildew, leaf spot, bacterial angular leaf spot, spot diseaseand anthracnose, 6 common diseases of cucumber.2. Studied the preprocessing methods of disease Images, including monochrometransformation processing, smooth handing, enhancement processing and edge detection andso on.3. Primarily extracted 14 characteristic value from monochrome statistic, color, andgeometry shape on processed disease image, and then use genetic algorithm to optimize andselect these characteristic value, finally got 8 representative characteristic parameter.4. Tried to use fuzzy pattern recognition to classify the cucumber disease image, studiedmembership principle identification and closest principle identification, designedcorresponding classifier, and built disease diagnostic model.5. Taking VC++ as the develop tool, written the system of cucumber disease imageprocessing and identify, finally realize the processing and identify of cucumber commondisease.Using fuzzy pattern recognition to classify the cucumber disease is a new attempt.Human thinking activities is a kind of fuzzy reasoning process, human identifying image isbetter than computer at certain aspects just because they can reason base fuzzy logic. Fornature plant, the disease occurrence parts were diverse in shape; using fuzzy recognition can classify it better.Adopting scientific and reasonable study methods for image collection, image preprocess,characteristic extraction and pattern recognition, we got ideal effect, the average recognitionrate for 6 disease is up to 93.3ï¼…, and it is evidently higher than naked eye recognition.