Development Of A Portable Nondestructive Instrument Based On Laser Backscattering Imaging To Detect Soluble Solids Content And Firmness Of Fruits

China is a big country of fruit production,but not a trade power.The main reason is that the post-processing technology of fruit in China lag behind,especially for the grading technology.Nowadays,the main nondestructive testing equipments are based on machine vision technology and near infrared spectroscopy technology.However,the machine vision technology is more suitable for the detection of the appearance quality of fruit,while the near infrared spectroscopy technology can detect the internal quality,but usually with high price.This study developed a commercial grading device of fruit internal quality based on laser scattering image.This equipment is of great significance to improve the modernization and automation of postharvest fruit classification,and to improve the economic value and competitiveness of China’s fruit in the international market.The contents ang results are as follows:1.The design and implementation of a portable systemWith the ARM embedded microprocessor and Linux operating system,C/C++ and Python language were used for programming.The selection of system hardware was determined,including semiconductor laser,camera lens,battery,development board,display screen,SD card and switch.A embedded software development platform was constructed,including the main control module,image processing module and prediction module,and embedded GUI development environment was built,with friendly interface and simple operation.2.The study on rapid detection of sugar and firmness of peachThe laser scattering images of peach(c.v.Hujingmilu and c.v.Zaobaihua)through the portable fruit internal quality grading system developed.The parameters of the system,such as camera resolution,exposure time,focal length,were determined.By the method of threshold segmentation,the color parameters(R,G,B,H,S,I)and texture feature parameters(pixel area,smoothness,three moments,consistency and entropy)were extracted from the image.It was found that the pixel area,the consistency and the entropy of the image parameters had good correlation with the SSC and F of the peach,and the PLS-DA and SVM-C models of SSC and F were constructed.The overall accuracy rate of PLS-DA models of SSC of two kinds of peaches were 82%and 90%respectively,while SVM-C models were 92%and 94%.The overall accuracy rate of PLS-DA models of F of two kinds of peaches were 88%and 90%respectively,while SVM-C models were 93%and 95%.The results show that the SVM-C model was better than the PLS-DA model for SSC and F of peach,and the SVM-C model was used to develop a portable fruit quality classification system.3.The study on rapid detection of sugar and firmness of appleTwo kinds of Fuji apples(c.v.Hongqian and c.v.Dashahe)were selected as the test object.It was found that the pixel area,the consistency and the entropy of the image parameters had good correlation with the SSC and F of the apple through the same research methods like peach.The overall accuracy rate of PLS-DA models of SSC of two kinds of apples were 82%and 88%respectively,while SVM-C models were 92%and 94%.The overall accuracy rate of PLS-DA models of F of two kinds of apples were 91%and 90%respectively,while SVM-C models were 96%and 98%.The results show that the SVM-C model was better than the PLS-DA model for SSC and F of apple,and the SVM-C model was used to develop a portable fruit quality classification system.