Laser Center Extraction Method of Circular Linear Structure Profile Measurement Based on Cluster Segmentation
Hole parts,such as bearings,cylinders,hydraulic cylinders,and artillery body tubes,are important to the field of aviation,spaceflight,auto,machine tool and heavy equipment manufacture industry.However,the inspection of inner surface quality of this kind of parts,especially deep hole inner surface,remains a difficult problem in the manufacturing field.More specifically,the difficulty of inspection system mainly lies in two aspects:firstly,the workspace is narrow,making it difficult to effectively use the existing inspection technologies;secondly,due to the limitations of the non-open space of the inner surface,it is easy to form blind zone during the task of measurement.To satisfied with the increasing requirements of quality inspection in above fields,some scholars have applied the optical method into the measurement system of hole parts because of its advantage of high accuracy,high robustness,rapid speed,non-contact,and low cost.However,when measuring inner surface of hole parts,the optical method often presents the problem of a limited or occluded field of view for the characteristics of hole parts.Due to the problem of occlusion,the well-known fringe projection profilometry cannot projected the fringe pattern to the inner surface and the desired deformed fringe pattern cannot be captured by a camera accordingly.So,circular structured light 3D measurement technology has been extensively employed to inspect accuracy of the hole parts,which is an active vision measurement technology based on the laser triangulation principle.Many works have been done to enhance the performance of circular structured light technique.However,occlusions are problematic because the light source and camera are placed on the same side of the measured object in the exiting technique.Additionally,the technique may be disturbed by speckle noise when metal surface is measured.Consequently,for structured light measurement technology,it is necessary to develop an efficient preprocess method to eliminate the disturbance to the most extent because the performance of the technology largely depends on high quality images.But this problem remains unsolved so far to our knowledge.In the work,the circular linear structure profile measurement method is proposed to inspect the quality of internal surfaces of hole parts.In our method,the light source and the image sensor are arranged on either side of the object to be measured to get a clear view during data acquisition.However,the proposed circular structure profile measurement technique is seriously affected by the quality of the extraction of the center of the laser center when a metal surface is measured.Concerned that removing the image of the light source used in the setup and laser speckle effect is vital to center extraction of circular structured light,in this paper,a clustered segmentation-based center extraction method is developed.The process of our method is as follows:The maximum inter class variance(OTSU)method is firstly used to adaptively segment the captured laser image into two parts:background and target image;then,the clustering method is used to separate the laser light source image and the laser speckle in the captured laser image;finally,sub-pixel laser center extraction is realized with help of the well-known Steger algorithm.Both simulation and experiment validate the correctness of the proposed method.The experimental results show that the proposed method can well suppress the influence of the laser light source image and the laser speckle during the extraction of the center of the laser center when compared with the typical filtering methods,such as median filtering.In addition,the proposed method can fulfil the task of center extraction efficiently,stably,and reliably.Therefore,it provides an effective way of extraction of laser center for three-dimensional contour measurement of deep hole parts.In the future work,we will focus on the development of accurate calibration method to convert 2D light stripes information into 3D coordinate information.
Deep hole measurementCircular line structured lightSpeckleCluster segmentationLaser center extraction
万方数据
维普
