Research Of Parts Size Measurement System Based On Vision Technology

Automatic identification and size measurement in industrial parts is one of the important means of production automation and intelligent.However,the most implemented measurement systems are simply measuring of a certain class of inerratic parts;there is no integration of multiple types of parts size measurement algorithms for applications and research.To solve this problem,a parts size measurement system is necessary based on vision technology.In this paper,there designs a size measurement system based on vision technology,and system is divided into three parts besides image acquisition module,defect detection module and sizing module.It incorporates defect detection and size measurement on a variety of parts,and achieves defect detection and sizing on a variety of types of parts in all positions.The experimental analysis shows that this system proposed measurement algorithm based on the edge information and improved Hough circle detection methods,then using OpenCV and C++language research on the size measurement algorithm on a variety of types of parts.The main research work and innovations are summarized as follows:(1)Defect detection algorithm based on template matching.On the basis of template built on research shape matching algorithm based on contour feature.Seen in pairs geometric histogram matching method execution speed,high accuracy template matching.Then given defect feature extraction and recognition algorithm,the surface marker and eliminate substandard parts,module size measurement to avoid measurement of the defective parts,eliminating part of invalid measurements,improve the detection efficiency.(2)Based on the size of the edge detection measurement algorithms.Advantages and disadvantages of external Rectangle and corner point method of rectangular and polygonal part size measurement is given based on the edge information applicable to any part of rectangular polygon size measurement methods;analysis and Hough transform circle detection Hough transform circle detection point method,an improved point Hough transform method,in which the point group chosen method improvements,making circle detection process more stable,more accurate measurements.(3)The measuring accuracy of optimization method.This paper optimized the proportion and the implementation process of calibration,you can use different calibration algorithm based onthe different parts of the characteristics,and it improves the accuracy of the calibration.Data compensation is realized by dichotomy,to simplify the calculation complexity.Methods of measurement data comparison shows that the improved measurement precision.(4)Design and realization of vision dimensional measurement system.System architecture design and build the system hardware environment,select the background lighting solutions,diffuse light and Genie camera,to capture clear images with high contrast parts;design and implement the various modules and system software flow control,to achieve the four types,eight kinds of parts dimensional measurement.Experiments show that the system can better complete parts size measurement,and operating speed is within 500 ms,measured parts are between 10 mm ~ 64 mm in size,absolute measurement error is within 0.1 mm,and it conforms to research requirements of the measurement.