Research On Representation Of Cloud Models For Uncertain Knowledge Of Wheat Scab

Uncertainty artificial intelligence is a hot topic in the field of computer research.Many machine learning algorithms,intelligent predictive tools and applications are also constantly innovating.In particular,the uncertainty concept,such as “around”,“about” or “approximately” appearing in intelligent research has attacked frequently in the research of various fields,and has also made part of the research focus on the study of uncertainty.In order to allow these uncertainties to be released,many methods of expressing uncertainty is developing.Wheat yield among the top three crops in various crops is an important source of food.Wheat scab is a kind of disease that has a wide scope and degree of damage in wheat diseases.The judging index of the serious condition is determined by the range of the panicles disease rate in the technical specification standard of wheat scab-matism [GBT15796-2011].Due to the hard division,there will be a situation where “two infinitely close critical values are divided into different levels”,it is difficult to define their grades for these critical points.Therefore,the reasonable expression of the classification system uncertainty in wheat head blight has certain guiding significance for water application.This paper takes the uncertainty of wheat scab in the classification diagnosis as the research object,and uses the cloud model as the tool of uncertainty knowledge representation to realize the uncertainty knowledge representation of the process of scab level division.Comprehensive evaluation and analysis of the results are provided.The main research work of this thesis is as follows:(1)Discuss the representation method of uncertainty knowledge,and the application of cloud model in various fields.This paper discussed some representations of uncertainty knowledge and four representation methods of uncertainty knowledge.Analyze the characteristics and appropriate application scenarios of each uncertainty knowledge representation method.Describe relevant theory and important concepts of cloud model.(2)Researched image segmentation methods for wheat scab disease.Since the disease area of wheat scab differs from disease-free,after detailed comparison of the advantages of different color space,the image is denoised in a color space that the light intensity has a low influence on the image quality.Then,extracted the diseased area of wheat scab.Under the color components in different color spaces,different segmentation algorithms are used to achieve the segmentation of the diseased area.(3)The color characteristic parameters of wheat scab were constructed and a color feature cloud model was generated.After the image segmentation operation,we extracte the disease area and calculate the values of color moment of the components about the three levels.Due to lacks persuasiveness if only use color moments as the attribute element,28 new color features are constructed,together with color moments to be color features.By regression analysis of the correlation coefficient between disease rate and color feature,the frequency of appearance of color parameters in each region was counted,frequency curves were drawn,feature values of color feature clouds were obtained according to the cloud transformation algorithm,and corresponding cloud models were generated..(4)A comprehensive evaluation model of the color characteristics cloud model of wheat scab was proposed.From the "3En" norm of cloud model,the contribution of cloud droplets to the concept was found to be concentrated in a certain range.According to this property,a color feature cloud comprehensive evaluation model was proposed,and finally the results of the evaluation are explained and analyzed.