Detection Of Apple Mildew Heart Disease Based On Functional Data Interval Selection

Apple is one of the main kinds of fruits produced in China,and its yield and cultivated area rank the first in the world.However,the low rates of excellent fruits and high-quality fruits have been restricting the development of apple industry in China,and mildew heart disease is one of the main reasons.If undetected,fruits with mildew core will be mixed with healthy fruits and flow into the consumer market,not only the health of consumers will be damaged,but also the reputation of merchants will be affected.Therefore,it is of great significance to control the flow of infected fruits into the market from the source.This thesis starts from the perspective of apple nondestructive testing,then use the idea of functional data interval hypothesis testing to provide theoretical guidance for interval screening of apple near infrared spectrum data.Then SVM algorithm combined with weighted subspace algorithm based on random subspace algorithm is adopted to classify spectral data,so as to achieve the purpose of detection of apple mildew heart disease.The classification accuracy on real data set of the proposed WS-SVM algorithm is 93.54%,which is efficient,convenient and accurate in the detection of apple mildew heart disease.