Cherries Defect Detection Based On Machine Vision

Machine vision is a fast,non-destructive and efficient detection technology that replaces human eyes with machines.It has been applied to many different industries.Its speed and accuracy provide people with an alternative to manual,automated,non-destructive and cost-effective technology to meet the growing production and quality requirements of multiple industries.Machine vision detection technology is widely used in modem agricultural industry,especially in the external detection of fruits.In this paper,machine vision technology and Lab VIEW software are used to extract the surface image information of cherries,and the corresponding image processing algorithm is proposed based on the information obtained,so as to realize the recognition of cherries surface defects.Then,according to this process,an automatic detection program of cherries surface defects was developed to realize the automatic detection of cherries defects in slow motion.This experiment laid a foundation for the next step of the system to complete automatic,real-time and fast cherries defect detection,and for the development of defective fruit screening and grading system.The main contents and results of this experimental study are as follows:1.The application of machine vision system in agricultural industry at home and abroad is briefly introduced.The research methods and effect of fruit surface defect detection technology at home and abroad are analyzed.2.Determine the design requirements of machine vision system in Cherries defect detection system,and select the hardware needed according to the experimental requirements,such as industrial camera,lens light source,light box,industrial computer,etc.3.According to the hardware parameters and experimental conditions,the appropriate image preprocessing operation is designed,the surface information of cherries image and background image is analyzed,various methods used in image processing are understood,and the segmentation basis and method of cherries and conveyor belt background image are determined.4.The feature extraction method of cherries surface defect is studied,and a new method of cherries surface defect recognition is designed,and the method of mask overlay is proposed to solve the problem of cherries surface reflection.5.According to the image processing method,the corresponding LabVIEW program was compiled,and the comprehensive correct rate of the software for cherries defect detection under moving condition was 83.4%,which basically met the experimental requirements.The software realizes the automatic detection and recognition of cherries defects in motion state,and adds the detection program of the center of mass position of the defective cherries to prepare for the subsequent construction of a complete defective cherries screening system.