Research And Implementation Of Organophosphorus Pesticide Residue Detection System Based On Machine Vision

In recent years,with the extensive use of pesticides,the problem of pesticide pollution has become increasingly serious,and the ecological environment in China has been severely damaged.In order to increase the supervision and management of pesticide pollution,it is imperative to develop on-site rapid detection equipment for pesticide residues.In this paper,based on the rapid on-site detection requirements of pesticide residues,an organophosphorus pesticide residue detection system based on machine vision was designed and its key technologies were analyzed and studied in depth.The specific research contents are as follows:(1)Based on the principle of pesticide enzyme detection card enzyme inhibition reaction and the characteristics of the image itself,the structural design of the machine vision acquisition device was completed,including: the selection of industrial cameras,the selection of industrial camera lenses,and the selection of light sources,etc.This completes the construction of the machine vision platform.(2)Using the established machine vision platform to obtain color images of pesticide residue detection cards,and study the probability density of six kinds of noises: Gaussian noise,Ruili noise,gamma noise,exponential noise,uniform distribution noise,and salt and salt noise.It shows that the noise of the pesticide residue detection card is salt and pepper noise,and then the noise of the detection card is filtered and denoised using digital image processing technology to complete the pretreatment of the detection card image.(3)Divide the target area of the pesticide residue detection card and compare the edge detection effects of Sobel operator,Roberts operator,Log operator,Laplacian operator and Canny operator.The results show that Canny operator The edge detection of the detection card is best,and the extraction of the target area of the pesticide residue detection card is performed using the Hough transform.(4)In order to explore the specific relationship between image information and pesticide residue concentration,this paper adopts RGB color space model to transform the color information of the detection card target area into R,G,B three-channel pixel values with digital signal characteristics.The pesticide residue prediction model was established by using particle swarm optimization and genetic simulated annealing algorithm.Combined with the characteristics of the detection card,the pixel values of R and B channels were used as the input of the prediction model,and the pesticide residue concentration was used as the output.By analyzing the dominance of the two prediction models,it is found that the concentration prediction model based on genetic simulated annealing algorithm is more suitable for concentration prediction.(5)Establishing Man-Machine Interactive Interface of Organophosphorus Pesticide Residue Detection System by Using Matlab GUI and Completing the Development of Pesticide Residue Testing Equipment Prototype.