Development Of Non-contact Detection System For Seamless Steel Pipe Surface Defects

Nowadays,China is one of the world’s largest producers and consumers of seamless steel pipes,widely used in various industries.However,surface defects may occur,significantly affecting the performance and safety of finished products due to the influence of material quality,production technology,environmental conditions,and other factors during the seamless steel pipe manufacturing process.With the rapid development of the economy and the improvement of industrialization degree,the demand for high-quality seamless steel pipe continues to increase.How to efficiently and accurately detect surface defects to meet the production quality standards of enterprises has become an urgent problem for many seamless steel pipe manufacturers.To address the problem above,a seamless-steel-pipe-surface defect detection system integrating image processing technology is designed and developed based on the multi-line structured light detection principle.Compared with the traditional detection methods(such as manual visual inspection,eddy current inspection,ultrasonic inspection,and magnetic particle detection),this system has many advantages,such as high precision,fast speed,convenient installation and deployment,and low cost.The main contents of this article are as follows:(1)Establishment of structural light detection model and hardware system design and construction.The imaging model and distortion model of the camera are analyzed and compared,and the detection model based on multi-line structured light is determined.Based on this model,the hardware system is designed and built for the surface defect detection of seamless steel pipe,and the actual deployment carries out on enterprises’ production lines.(2)System calibration and extraction of dense light fringe center.A variety of camera calibration and optical plane calibration methods are analyzed and compared.According to the actual requirements of this paper,the calibration method of Zhang Zhengyou is determined by camera calibration.A multi-line structure cursor determination method based on free-moving plane targets is proposed in optical plane calibration,which simplifies the target demand and calibration process and dramatically reduces the difficulty of multi-line structure cursor determination.It can calibrate 200 optical planes simultaneously.In addition,the typical optical fringe center extraction algorithm is analyzed and compared,and the gray barycenter method is improved under the standard of focusing on real-time accuracy.A multi-fringe center extraction method with specific adaptability is proposed,which can accurately extract dense optical fringe centers without losing real-time performance.(3)Realization of surface defect detection and system software.The reverse vision method simplifies the center points of the extracted optical fringe,and the suspected defect points are retained.A three-dimensional point cloud generates the suspected defect points.The alleged defect points are finally judged by the relationship between the distance between these points and the axis of the steel pipe and the standard pipe diameter of the steel pipe.In this way,the time of the system is reduced by 84.6% while the detection accuracy is maintained.The detection accuracy can reach 93.2%.In addition,according to the needs of enterprise users,a supporting software system is realized.Through the experimental evaluation,the detection speed of the whole system can reach 1.0m/s,while the sampling interval in the lateral direction can reach 1.0mm.