Design Of Performance Testing System For Seed Metering Of Dibbler Based On Machine Vision

Cotton is the pillar industry of Xinjiang’s economic development.The precision level of the seeding device affects the output of cotton.One of the most important methods of precision seeding is hole sowing,and the most important part of the seeder is the hole seeder.Therefore,before the planter leaves the factory,it is necessary to appraise and test the performance of the hole planter.This article summarizes and analyzes the four methods of seeding performance testing of the hole planter,including manual detection methods,photoelectric sensor detection methods,high-speed photography detection methods and machine vision detection methods.The advantages and disadvantages of the four detection methods are proposed and the reliance is pointed out.The machine vision method to detect the seeding performance of the hole planter is still the trend of future development.To this end,a machine vision-based precision metering performance detection technology for the hole planter is proposed.Lab VIEW image processing technology is used to process the image of the cotton seed,so as to realize the precision detection of the seeding amount of the cotton seed per hole and determine the hole of the hole planter.Rate,replay rate and pass rate,and then judge the seeding performance of the hole seeder.The main research contents and analysis results are as follows:(1)Build a hardware image acquisition system,which includes industrial cameras,lenses,light sources,light source cassettes and data acquisition cards,etc.,compare and study the types of each device,and build a hardware system that is most suitable for the performance testing of the hole planter.After the hardware image acquisition equipment is built,use the Grab control in Lab VIEW to acquire images continuously at a high speed,and perform masking operations on the acquired images.(2)The image processed by the image mask operation is processed by the color model of the image(RGB and HSL),and the cotton color and the background color are compared,and the two color components with the best effect are selected as(R and L).(3)Perform image gray-scale operation processing,adjust the brightness of the image and enhance the contrast of the image.After the image gray-scale processing,the position of the cotton seeds and the number of cotton seeds in each hole can be seen;Sobel algorithm is used The edge detection is performed on the cotton seed image,and the two-dimensional contour of the cotton seed is recognized.(4)The brightness and filtering of the image have significantly improved the fidelity of the image quality.The image filtering process is performed by the formula method.After the noise reduction process,the position of the cotton seed and the number of the cotton seed can be clearly seen.It shows that the result of image processing has strong feasibility.(5)Finally,the experimental data is analyzed.The error between the actual value and the measured value of the total hole seeding amount is 0.5%-6.0%,and the total hole seeding amount can meet the actual work.The actual qualified rate of the number of holes is 93.5%-98.0%,the replay rate is 1.0%-3.5%,the hole rate is 1.0%-3.5%,the qualified rate of the number of holes detected by the system is 91.5%-97.0%,and the replay rate is 2.0 %-4.5%,the cavity rate is 1.0%-4.0%,the maximum relative error of the hole number pass rate detection is 2.7%,the detection accuracy can meet the actual needs,and the experimental data analysis is within the error range,indicating that the machine vision-based The seeding performance detection system of the hole planter is feasible.