Abstract:It has great significance that using the machine vision technology to detect the quality of postharvest litchi fruit. Firstly, the camera and fluorescence spectrometer were used for the spectrum analysis of litchi image, the emission spectrum characteristics were analyzed under the fluorescence as excitation light, which determines the feasibility of the visual detection method of litchi fruits with different fluorescence exposures. Then, the machine vision system of different light switch controls were designed, the red, blue and green fluorescent lamp were selected, and the singlechip microcomputer system was used to control the switch of the LED lamps, of which the interval is 1s; meanwhile, the image acquisition system triggered the camera to take images, the frequency of the light switch in keeping with the number of taking image times. The grey level histogram of the fluorescence image for normal and microdamaged state of two kinds of litchi fruit was statistic analyzed, the image recognition method for the micro damaged litchi fruit was determined by using blue fluorescent as light source and the V component of HSV color space. Then the exploratory analysis was used for the statistics and analysis on test results of litchi fruit visual inspection. The grayscale image segmentation threshold of the normal and microdamaged litchi fruit was determined. The grayscale image threshold segmentation, the morphology processing and the optimized Hough circle fitting method were used to the litchi images, which realized the design of the machine vision intelligent classification system for litchi fruit. The test results show that: the recognition accuracy of the normal and microdamaged litchi fruit is 92%, which can provide technical support to intelligent detection technology for postharvest fruit and vegetable.