Evolutionary Algorithm For Automatic Defect Detection On Vehicle Body Paint

Cars have become an indispensable means of transportation in people’s daily lives.With the increase in car usage,car safety issues also become more important.As a part of the safety of the car body,the paint film on the car body can effectively isolate external water vapor and prevent oxidation.The traditional defect detection of car body paint film is to track the workshop assembly line by human eyes to determine whether the car body surface paint film contains defects,and the subsequent manual re inspection is needed.Affected by the workshop environment,human eye resolution,human factors and other aspects,human eyes are prone to visual fatigue,which makes the detection become inefficient,inaccurate and increases the cost.This paper describes the importance of body paint film,and introduces the research status of body paint film defect detection at home and abroad.Compared with the existing research methods,evolutionary algorithm is proposed to solve the disadvantages of low accuracy and long time-consuming of traditional methods.On the basis of ant colony algorithm and particle swarm algorithm,combined with the characteristics of car paint image pixel gradient and gray threshold,an ant colony defect detection algorithm and particle swarm defect detection algorithm for car body paint film defect detection are proposed.To improve the adaptability of the two algorithms,a defect detection algorithm based on particle swarm optimization and ant colony optimization is proposed,which effectively improves the efficiency and accuracy of defect detection in car paint image.The research contents of this paper are as follows:1)Because the defect detection process of car body paint film is in a high-intensity light environment,the image collected by the camera will appear local reflection phenomenon,which will cause interference to the detection of car body paint film defects.In this paper,a reflective area elimination algorithm based on body paint film is proposed.The highlight pixels are extracted by using the color attribute of RGB channel,and the reflective area is eliminated by combining the pixel gradient and gray threshold.2)Based on the traditional ant colony algorithm and particle swarm optimization algorithm,the edge detection method of ant colony paint film defect detection is proposed.Combined with the characteristics of bright,reflective and turbid car body paint film,a novel position updating method and random selection strategy are proposed for particle swarm optimization algorithm.Aiming at ant colony algorithm,a new pheromone updating algorithm,the random selection method and the decision rules of image edge are designed and implemented to realize the accurate detection of the defect location.3)In order to solve the problems of premature convergence and longtime cycle of the algorithm,the ant colony algorithm and particle swarm optimization algorithm are combined.The particle swarm optimization algorithm is used to reduce the blind search in the early stage of ant colony algorithm,and the particle swarm optimization algorithm is used to improve the detection accuracy of particle swarm optimization,and the edge of the paint film defect of the vehicle body is detected accurately.The experimental results show that the method proposed in this paper can detect the defect position in the image effectively,accelerate the detection speed,improve the shortcomings of manual detection,and greatly improve the detection efficiency.