Detection And Experimental Analysis Of Potato Diseases Based On Near Infrared Spectroscopy

The implementation of potato staple grain strategy has accelerated its industrial development,and potato quality directly affects the economic benefits of deep processing industry,which has become a topic of increasing concern.Black heart disease and ring rot are common internal defects of potatoes.Once they appear,they will seriously affect the quality of potatoes and even the health of consumers.In order to improve potato quality and accurately identify potato internal defects,a potato disease detection and sorting device based on near infrared spectroscopy(band range 900 nm～1700 nm)was designed in this paper,which can simultaneously detect and remove two internal defects.The main contents and conclusions of this paper are as follows:(1)A potato defect detection and sorting device based on near infrared spectroscopy technology and electrical linkage was designed.The device consists of a miniature near infrared spectrometer,a conveying system,a culling system and a PLC control system.The micro-NIR spectrometer can realize the non-destructive detection of potato internal defects.Conveying system realizes potato conveying by electric roller and belt;The elimination system uses pneumatic components to eliminate defective potatoes.PLC control system controls the operation of the device to complete the detection and sorting of defective potatoes.(2)The identification model of potato black heart disease and ring rot was established based on near infrared spectroscopy.Healthy potatoes,black heart potatoes,and ring rot potatoes were regarded as the research subject.Through first-order derivative,smoothing,Multiple Scattering Correction(MSC)and Standard Normal Transformation,The original spectrum of potato was pretreated by four pretreatment methods.Principal Component Analysis(PCA)and Soft Independent Modeling of Class Analogy are established.SIMCA,Support Vector Machine(SVM)potato discrimination model.By comparison,MSCSIMCA potato discrimination model was confirmed to have the best discrimination effect on black core potato and ring rot potato,and the recognition rate reached 94.89% and91.01%,respectively.(3)A potato optimal discriminant model based on characteristic wavelength was established.Using Competitive Adaptive Reweighting Algorithm,Feature wavelengths were extracted by CARS,SPA and Uninformative Variable Elimination(UVE).The optimized MSC-SIMCA discriminant model was established by extracting 15,9 and 13 characteristic wavelengths by three methods.The results showed that UVE-MSC-SIMCA had the best recognition effect on black core potato and ring rot potato,and the recognition rate reached 97.08% and 96.63%,respectively.Through optimization,the accuracy of the model is improved and the detection speed is accelerated.(4)The working conditions of the separation device were studied by central response surface test and the optimal working parameters of the separation device were determined.In order to reduce the potato in the process of sorting falling injury,explore the sorting device working conditions of potato fall injuries and potato best separation condition,the influence of separation device of conveying speed and carved output,the drop height and the moisture content of potato is test factors,through the single factor experiment and the analysis of the center response surface test the influence of the above factors on potato fall injury.The results show that it is feasible to detect and separate potatoes when the conveying speed of the device is 0.3 m/s and the picking power of the sorting device is15 N.Moreover,when the potato dropped to 45 cm height and the potato moisture content was 75.61% after 5 days of potato placement,the sorting device had the least damage to potatoes.The model accuracy and stability of the device were tested with 23 healthy potatoes,20 black potatoes and 17 ring rot potatoes that were not involved in the modeling.The results showed that when the conveying speed was 0.3m/s and the removal force was15 N,the device could detect both black and ring rotten tubers simultaneously,and the recognition rate of black and ring rotten tubers was 95.02% and 94.11%,respectively.The potato damage after sorting was less than grade 1.