Research On Apple Hardness Portable Device Based On Near Infarared Spectroscopy

Apple is popular with consumers because of its crisp,sweet,rich vitamins and acids.With the continuous improvement of the quality of life,consumers are increasingly demanding the quality of apple.Hardness,as one of the important indicators of apple's internal quality,plays an important role in the quality of apple.Therefore,predicting the hardness of apples during the apple production process and determining the optimal harvesting period will help increase economic efficiency.Traditional apple hardness testing mainly relies on portable hardness testers and texture analyzers for loss detection.Near-red spectrum detection technology has the characteristics of high efficiency,no damage and easy operation,and is widely used in apple quality detection.At present,there are less equipment for non-destructive testing of apple hardness.In view of the above problems,this paper proposes a research on apple hardness portable devices based on near-infrared spectroscopy.Extracting characteristic wavebands of near infrared spectra using random frog leaping algorithm.Then,combining multiple linear regression(MLR)and support vector machine(SVM)algorithm,constructing the near infrared spectra of apple firmness prediction model.The spectrum acquisition module with characteristic band LED light source and photoelectric sensor was designed to achieve nondestructive testing of apple hardness.Apple hardness non-destructive testing equipment is of great significance for the prediction of the harvesting period and the quality inspection in apple production.The main work and results of this article are as follows:(1)Construction of apple hardness nondestructive testing platform and extraction of characteristic bands.Analysis found that the hardness value of apple matured first increased and then decreased,and it was closely related to the content of pectin and cellulose.Construction of apple hardness near infrared spectroscopy diffuse reflectance detection platform with using HL-2000 tungsten halogen lamp,ocean optics OFS-1100 spectrograph and computer.Selecting 3 points at equal intervals of 120°in the direction of the apple equator and repeat 3 times to collect apple spectral data with a wavelength range of400-1100nm.Hardness data were detected after peeling the apple at the same position in the spectrum acquisition.Finally,the eight characteristic bands of the spectral data are extracted by the method of random frog leap,which are 589 nm,638 nm,707 nm,794 nm,832 nm,880nm,996 nm,and 1043 nm,respectively.(2)Research and design of apple hardened portable non-destructive testing equipment hardware system based on near-infrared spectroscopy and device appearance.According to the different functions of the hardware system,the hardware system is divided into a data processing module,a light source driving module,a spectrum acquisition module,and a user interaction module.The data processing module mainly includes S3C2440 processor,flash memory,system clock source and I/O interface.Design development of light source driving circuit through PT4115 driver chip.Spectral acquisition module is mainly composed of FDS1010 photoelectric conversion diode and ADS1115 AD conversion chip.User interaction module consists of TFT-LCD touch screen.Through Altium Designer design schematics and PCB(Printed circuit board).Finally,to complete the equipment design with using Proe software.(3)Development and design of apple hardness portable nondestructive testing equipment software system and human-computer interaction interface.According to the requirements for the development of embedded software systems,a software platform based on the Linux operating system was built,and a cross-compilation environment was established.Uboot is transplanted,the operating system kernel is transplanted and trimmed,the root file system is built,and the I2C driver based on ADS1115 is designed.Porting Qt5.6,and design a graphical user interface design based on TFT-LCD touch screen.(4)The apple hardness prediction model was built,and the modeling effects of different algorithms were verified and compared.The apple hardness nondestructive testing equipment embedded with intelligent algorithms was verified and tested.The apple hardness prediction model was established by MLR and SVM respectively.The study found that the MLR hardness prediction model's correlation coefficient(R~2)is 0.8168 and the root mean square error(RMSE)is 1.0370N.The SVM hardness prediction model's correlation coefficient(R~2)is 0.7022 and the root mean square error(RMSE)is 0.7086N,between which the MLR hardness prediction model has better effect.The equipment verification experiment results show that The determination coefficient of the apple hardness measurement and equipment prediction value is 0.7177,the mean square error is 2.459N,and the average error is 4.1334N,the standby time is 20 hours,and the continuous working time is 10 hours.