Study On Nondestructive Detection Of Apple Quality Based On Hyperspectral Imaging Technology

China’s fruit production and cultivated area ranks first in the world.Fruit species variety, its level of consumption has become increasingly growth.But the market for fruit detection technology compared with the foreign detection technology is relatively backward.Fruit quality detection technology is relatively backward.The development of the nondestructive testing technology is more and more become a hot trend, and the characteristics of hyperspectral technology has made it become the best in the nondestructive testing technology of detection means.It combines the advantages of image technology and optical technology, an image contains both the image information of the fruit also contains its spectral information, but for rapid detection of fruit quality of inside and outside.Apple is one of the main fruit of our country, much attention has been paid to its quality for consumers, this article will apple as the research object, using the compose a specular like technology for nondestructive testing of apple’s internal and external quality parameters is studied.Using the method of PLS and BP- ANN apple sugar, pulp hardness prediction model is set up;For external minor damage detection, respectively using the feature band principal component analysis(PCA) algorithm and the lowest noise separation transformation(MNF) based on the image processing method to detect the apple surface slight injury.And compared to two kinds of detection algorithm, find apple detection algorithm.In this paper, the research results provide a reference for real-time on-line detection system.In this paper, the main work is as follows:First,Hyperspectral image system, spectral information were collected to after pre-processing, comparing various pretreatment methods. Multiple scattering correction method is used for pretreating the original reflect spectral data, it has a better performance for the prediction and identification model.Second,The establishment of PLS and BP-ANN apple sugar content and flesh firmness analysis model.the model of the sugar content and the flesh firmness of the apple built by the BP neural network is superior to the partial least square method.Third,Five wavebands are selected as the effective wavebands based on the coefficicnt curve of I-RELIEF method conducted on spectra extracted from intact and bruise surface.Fourth,The bruises detection algorithm is developed based on the effective wavebands and MNF transform method., using the algorithm in different time stages minor damage, the results with a 97. 1% overall detection are rate.