Research On Apple Maturity And Associated Quality Factors Based On Nondestructive Detection

China is a major production in apples.In 2015,the cultivation area and production of China's apples all ranked first in the world,but fresh apple exports amounted to 833,000 tons,accounting for only 1.9%of China's total apple production and its market share in the world and China.The status of apple producing countries is extremely inconsistent.As a key link in China's apple industry chain,harvest maturity has a direct impact on Apple's quality level,product value and economic benefits.Based on the analysis of apple maturity process,this paper proposes a non-destructive testing method based on the fusion of mature quality factors,and constructs a comprehensive forecasting model of maturity(IQI).Based on this,it uses factor analysis to extract the evaluation index(I)of apple's internal comprehensive quality maturity,and verified the prediction performance of both.The results show that the establishment of a comprehensive evaluation index of maturity not only expounds the interrelationships among the quality parameters,but also reveals the gradual progress of the comprehensive quality information of apple during its maturity.It also has implications for the quantitative determination of apple maturity and is The construction of apple maturity evaluation standard system and the standardization of fruit harvest provide a theoretical basis.The main contents and conclusions of this study are as follows:(1)Research on internal quality inspection methods and construction of visible/near-infrared spectrum test platform.Based on the analysis of apple maturity mechanism,the main correlation factors of non-destructive testing of maturity were extracted,and the feasibility of comprehensive prediction of maturity based on near-infrared diffuse reflectance spectroscopy and apple maturity related indicators was demonstrated.Finally,apple soluble solids content,pulp,were proposed.The hardness and apple shade parameters are the major predictors of maturity.For the requirement of apple diffuse reflection feature band selection,this paper constructed an apple maturity NIR spectrum acquisition platform to obtain the full spectrum of sample data,which provided the basis for the subsequent selection of feature bands and model establishment.(2)Research on predicting model of maturity related indicators.Based on the non-destructive testing platform of apple,the data of diffuse reflectance spectrum,soluble solids,pulp hardness,and luster of Fuji apple in Gansu Province were obtained.Feature wavelengths were extracted for different quality indicators.For CARS and RF algorithms,the characteristic wavelengths of the soluble solids model were 81 and 55,respectively;the characteristic wavelengths of the pulp hardness model were 49 and 64,respectively;and the color parameters L,C,and h model characteristic wavelengths.The numbers are 22 and 64,11 and 57,11 and 42 respectively.The best predictive model of soluble solids in apple was RF-PLSR,among which_PR is0.906,RMSEP is 0.744.The best predicting model of apple flesh hardness was RF-SVR,where _PR is 0.825,RMSEP is 0.99;the parameter L was the best model for predicting the flesh hardness.SVR,where _PR is 0.954,RMSEP is 1.06;parameter C,the best pulp hardness prediction model is CARS-SVR,where _PR is 0.942,RMSEP is 0.87;parameter h the best pulp hardness prediction model is RF-SVR,where _PR is 0.956,RMSEP is 2.455;(3)The establishment and validation of a comprehensive evaluation index forecasting model for apple maturity.Combined with the existing research,the above apple maturity related indicators were linearly combined to obtain the apple comprehensive maturity index IQI,the number of characteristic wavelengths extracted using RF to IQI was 76,and the best apple IQI prediction model was RF-PLSR._PR is 0.938,RMSEP is 0.216;using the factor analysis to obtain the internal quality index I of Apple,the number of feature wavelengths extracted using CARS for I was 13,and the best predictive model of apple I parameter was CARS-SVR,where _PR is 0.909,RMSEP is 0.259.The detection of single maturity index is not representative,and it is easy to cause errors in the identification of mature period.The establishment of the maturity detection model has achieved synchronous and comprehensive detection of multiple qualities,simplified the detection model,and improved the prediction ability and operation speed.