Research On Defect Detection And Sorting System Of Red Jujube Based On Machine Vision

Red Jujube is well known for its rich nutrition.Due to the influence of growing environment and picking technology,there are many defects on the surface of Red Jujube,such as surface damage,crack,etc.,in order to separate jujube with different defects,a defect detection and sorting system based on machine vision is studied in this paper.Firstly,the whole system is designed according to the sorting requirement,and the internal subsystems are studied theoretically.In this paper,the feeding device and the conveying device are designed,optimized,trial-produced and tested,the image acquisition method is improved,and the defect Detection Algorithm is compiled,design a targeted detection scheme.The main contents of this paper can lay a solid theoretical foundation for the future large-scale application of machine vision-based defect detection on Jujube surface.The specific research contents are as follows:The overall design of the detection system.According to the National Standard and the local enterprise standard,the types of surface defects of red dates were determined as damage and crack,and the sorting damage and crack of red dates were studied,the whole structure of the defect detection and sorting system based on machine vision is designed.The main structure of the system includes feeding device,conveying device,detection system and sorting system.Feeders and conveyors.To understand the types and basic working principle of the existing feeding device,summarized the shortage,according to the needs of this article to choose the appropriate feeding mode,through pre-test and theoretical calculation to determine the basic structure of the feeding device.The interaction between Jujube and material of feeding device was studied,and the rolling Friction Coefficient of Junzao Jujube on galvanized sheet was measured,the minimum value of the angle between the bottom of the hopper and the horizontal plane is determined according to the measured parameters such as the size,the density and the dynamic friction coefficient of the Jujube.The device is simulated with Edem 2018 software,and the internal structure and outlet size of the feeding device are optimized according to the simulation results.The trial-manufacture of the feeding device and the deployment of the stepper motor,controller and power control system are completed according to the design,and the performance test of the feeding device is carried out,the effect of each factor on the feed performance was studied at three levels.It was found that when the outlet was 44(cm),45(cm)and 46(cm),the single particle rate was 83.The two-factor interaction test was designed to explore the interaction of the drum rotating speed and the outlet size on the feed performance,and the quadratic polynomial was used to fit the results.Using Solid Works 2018 software to establish its three-dimensional model,confirm its parameter is 25050(cm),height is 70-100(cm)adjustable and its speed is linear adjustable between 0-8(m/s)and trial-manufacture the device.Defect detection system of Chinese date.The whole structure of the image acquisition device is designed,the three-dimensional model of the device is built by Solid Works 2018,and the frame of the adjustable camera is put forward according to the material characteristics of Junzao.According to the need of image acquisition,the brightness of Halogen,LED,high frequency fluorescent lamp,xenon lamp and incandescent lamp is selected as the system light source with adjustable brightness,long service life and low cost,select the appropriate camera according to the requirements of image quality and camera performance.The relationship between the object distance and the imaging size is analyzed.According to the imaging size,the optimum distance between the object and the Selected Camera Lens is analyzed.The Algorithm of defect detection is programmed.Visual Studio 2010 as the Development Platform,C language as the programming language,the use of OPENCV3.0 machine vision library and camera SDK tool kit to determine the defects of red dates.Sorting system.According to the current fruit sorting mode and the material characteristics of red dates,the whole scheme of red dates sorting system is determined by adopting the sorting mode of pneumatic Piston Push Rod,and according to the design scheme,the concrete parameters of the executive mechanism of red dates sorting are determined,using GX Developer 9.0 to write PLC control program,with PLC as the controller,to write a time-delay program through the control of solenoid valve reversing to complete the push rod expansion and contraction to realize the on-line sorting of different jujube defects.Finally,the sorting performance of single defect Jujube and multi-defect Jujube was tested in the whole sorting system,and the regularity of the sorting accuracy changing with the speed of conveyor belt was studied,the test results show that when the conveyor belt speed is 0.1 m/s,it has the best sorting effect,and the sorting correction accuracy of surface damage defect is 73.3%.