The Image Segmentation Of Potato Diseased Spots Based On Genetic Frog Leaping Neural Network

As grain,vegetable,feed crop and processed industrial material as well as biomass energy crop,potato of wide use has vigorous demand in the market and vast potential development.Boasting of short production circle,high yield potential,broad market demand and good economical benefit,potato has been widespread grown across the country for a long time.With the expansion of potato's planting scale,the increasingly serious disease and insect pests has become one of the main factors that hampers potato's yield and quality,which greatly affects potato's development.Therefore,utilizing computer vision to recognize diseased spots in the potato image plays an important role in both practical significance and application value.During the process of diseased spots image recognition of potato,the image segmentation of potato is the key step to extract the feature of potato's diseased spot image and recognize disease.This text mainly focus on the problem that segmentation result lacks of to some degree objectivity since utilizing traditional PCNN to carry out potato's diseased spot mage segmentation requires repeatedly manual operation in test and value.To solve the problem that parameter setting needs manual interference,genetic SFLA is put forward to improve image segmentation effect of PCNN.The main tasks of this text are as the following:(1)Combine genetic algorithm with SFLA to propose the genetic SFLA,to solve SFLA's low convergence speed and its easily plunging into local optimization.(2)Applying genetic SFLA to PCNN' parameter adjustment and build SFLA-PCNN model which doesn't need people to determine a parameter.(3)Use optimized neural network in colorful potato's spot image segmentation,and carry out the segmentation on potato's late blight,ring rot,gray mold and Spongospora subterranea six diseases.Then make a discussion and analysis on the segmentation result.The result shows that SFLA-PCNN model can effectively extract potato's disease from background area,which further lays good research foundation for potato's disease feature extraction and model recognition as well as potato's disease prevention and treatment.