Font Size: a A A

Research And Application Of Apple Sugar Detection Method Based On Near Infrared Technology

Posted on:2019-07-25Degree:MasterType:Thesis
Country:ChinaCandidate:D Y ChenFull Text:PDF
GTID:2351330542984563Subject:Pattern Recognition and Intelligent Systems
Abstract/Summary:PDF Full Text Request
Near infrared spectroscopy technology is more and more mature,more and more portable spectrometer.But there are still some problems such as high price and complicated operation.Miniaturization,intelligence and cheapness are the development trend of portable detector.Therefore,it is very important to propose a new apple sugar detection method to realize the miniaturization,low cost,and intelligence of the detector.In this paper,a method for detecting apple sugar degree based on multiple feature wavelengths is proposed,and the overall scheme of the detection device based on this method is designed,and the optical structure design,circuit design and software program development of the detection instrument are expounded.The cheapness and miniaturization of the detection instruments have been realized based on cheap LED light source,unique optical structure design.The designed near infrared detection instrument is used to obtain the diffuse reflection signal absorbance with the light of830nm,880nm,940nm(characteristic wavelength)and 850nm(reference wavelength),and the multiple linear regression(MLR)and partial least squares(PLS)forecast model are established.By comparing the parameters of the model,we can see that the multiple linear regression model parameters(R~2=0.9092)are superior to the one-variable linear regression model parameters(R~2=0.8367).The predictive correlation coefficient(R_p)and the predicted root mean square error(RMSEP)of partial least squares are 0.8957and 0.4052,respectively.Although the prediction accuracy of the model based on the spectrometer apple sugar detector(R_p=0.924>0.8957,RMSEP=0.334<0.4052)is better than the prediction accuracy of the experimental model,but its cost is far higher than the device.The experimental results show that the detection device based on multi-feature wavelength basically meets the requirement of apple sugar detection accuracy.The detection method of near-infrared apple sugar content based on multi-characteristic wavelength proposed in this paper is feasible,and it is better than apple sugar content detection method based on single characteristic wavelength.At the same time,this research can provide reference for the design and development of practical apple sugar content detection instruments.
Keywords/Search Tags:Non-destructive testing, apple sugar content, multiple characteristic wavelengths, diffuse reflection, photoelectric detection
PDF Full Text Request
Related items