Research On Geometric Parameters Measurement System Of Riveted Plates Based On Machine Vision

Thin sheet parts are widely used in automobile manufacture,electrical products,precision instruments and other industrial fields,and this kind of parts have the characteristics of thin thickness,constant change of hole's shape and hole's distribution position,the measurement and control of machining quality is an important production link.Combined with the practical application,in this paper,a visual measurement system for identifying the shape of distributed holes on riveted thin plates and measuring their geometric parameters is studied,it can provide technical support for the automatic transformation of rivet machine's supporting devices,while measuring geometric parameters of rivet sheets.According to the requirement and based on monocular machine vision technology,the design of rivet thin plate geometric parameter measurement system is completed.The vision system for obtaining images is designed and built based on the improved rivet machine automatic feeding device,including the selection of industrial camera,lens,image capture card,the design of illumination mode and the background of the plates clamping platform,and calibration of vision system;Image processing as the core of this paper,including image preprocessing,image threshold segmentation,feature extraction,hole recognition and parameter extraction algorithm of geometric parameters,first of all,the image interference is analyzed and filtered in the process of preprocessing,then threshold segmentation and feature extraction are completed with the rule of image gray distribution,and single threshold image segmentation method including gray histogram method iterative method,Otsu method are studied.According to the efficiency of the segmented image,the Otsu method is applied to segment the image.In view of the segmentation error of dividing the background into targets in the process of segmentation,an ROI extraction algorithm based on the principle of image connectivity is proposed,finally,the recognition of holes is realized on the basis of the characteristics of geometric holes in polar coordinates,the geometric parameters of riveted thin plates are measured by using the method of gravity method and the area geometric relations of different holes.By processing the image of riveted plates with different shapes and sizes with the algorithm which has been certified,has verified the applicability of the image segmentation,hole recognition,hole geometric parameter extraction algorithms used in this paper,and the factors that affect the processing results are analyzed,combined with the environment used in the measurement system,solutions and improvement schemes are put forward.