Research On High Precision Measurement Of Shape And Sizes Of Complex Workpieces Based On Machine Vision

With the rapid development of high-end CNC machine tool technology,the structure of workpiece products tends to be complicated and integrated,and the requirements for detecting and controlling the accuracy,quality and appearance defects of workpieces are becoming higher and higher.At present,the traditional detection methods for dimensional tolerances,shape parameters and surface defects of complex workpieces are mainly based on human eye observations and local single-point contact measurement.The influence of human factors is large and the detection efficiency is low.Therefore,how to achieve high-quality,high-efficiency,low-cost comprehensive detection of complex workpieces is an important key technology.In this paper,the piston of the automobile engine is taken as the research object.With the aim of high-precision non-contact automation and intelligent detection,computer vision technology is used to study the detection of key indicators such as shape,size and defect of complex workpieces.The main contents are:First,propose a dual camera calibration measurement method.Technical analysis of high-precision measurement of workpieces,taking into account the optical limit problems faced by high-precision measurement requirements,by analyzing the resolution,lens,working distance,depth of field,etc.,designing a reasonable measurement method to minimize the measurement process The error caused by the inevitable movement.Second,design complex workpiece high-precision inspection station.According to the specific content of the 19 detection items,the overall inspection process is divided into four inspection stations,and the camera and the reasonable position of the light source are designed.At present,a simulation test station combined with a camera and a light source is initially established for dimensional measurement research.Third,two kinds of detection methods are proposed for the diameter measurement of the top surface of the piston.For the case where the piston diameter is less than 5 cm,a local binarization edge detection method based on the principle of coarse to fine is adopted.Themethod combines the characteristics of the top surface image of the piston,taking into account the fact that the reflection of the workpiece itself is serious,resulting in uneven distribution of grayscale,and directly performing the edge detection effect is not ideal,which is obtained by the method of edge detection from coarse to fine local binarization.Precise edge.For the case where the piston diameter is larger than 5 cm,a sub-pixel fitting detection method based on the sigmoid function from the coarse to the fine principle is employed.The method solves the optical limit problem that the large-size piston cannot meet the high-precision measurement requirement,and the double-camera calibration measurement method further improves the accuracy.