Research On A Binocular Structured Light Based Refractory Brick Dimensional Measurement And Defect Detection System

The quality of refractory bricks has a significant impact on the service life and production safety of high-temperature equipment such as kilns,boilers,and smelting furnaces.Before being boxed for shipment,the manufacturer performs strict size measurement and defect detection on the refractory bricks.Currently,refractory brick inspection is mainly done manually,which is highly subjective and inefficient,and cannot meet the development requirements of enterprises.With the development of machine vision and 3D imaging technology,the use of 3D point clouds for refractory brick inspection is becoming increasingly mature.This article studies a refractory brick size measurement and defect detection system based on binocular structured light,with the following main contents:(1)The scheme design of the refractory brick detection system has been completed.Based on the characteristics of refractory bricks,the 3D camera is selected to ensure the high accuracy of the collected point cloud data.The PLC is selected according to the process flow of the detection system.The point cloud processing algorithm software platform is selected based on the detection requirements,and the software framework of the upper computer is determined.(2)The refractory brick point cloud preprocessing algorithm is studied.For the point cloud data in the non-target area,a suitable parameter point cloud pass-through filtering algorithm is used to remove it.For the scattered points around the edge of the refractory brick and the reference surface,two methods of statistical filtering and radius filtering are designed and analyzed,and the statistical filtering with better robustness is selected to remove the scattered points.In order to segment the refractory brick point cloud and the reference surface point cloud,an automatic threshold point cloud segmentation algorithm based on K-means clustering is proposed.The segmentation threshold is calculated by K-means clustering,and the refractory brick point cloud and the reference surface point cloud are segmented using this threshold.(3)The refractory brick size measurement and defect detection algorithm is studied.Based on the high height difference between the defect point cloud and the surface point cloud of the refractory brick,a refractory brick defect detection method based on tilt correction is proposed.Tilt correction is performed by calculating the pose of the refractory brick point cloud and performing rigid transformation,and then the defect is filtered based on the Z value and the defect threshold.The refractory brick point cloud has a certain regularity,and the length and width of the refractory brick are obtained by fitting the OBB bounding box to the refractory brick point cloud.The refractory brick point cloud and the reference surface point cloud are fitted with a least-squares method to obtain the distance from the fitting plane to the origin of the coordinate system,and then the two distances are subtracted to obtain the height of the refractory brick.The test results show that the average errors of the refractory brick length,width,and height are 0.31 mm,0.307 mm,and 0.235 mm,respectively,and the accuracy of defect detection is 97.25%.The detection time of a single refractory brick is 970 ms,which meets the actual onsite detection requirements.