An Area Measurement Method Based On Machine Vision Detection Technology For Cable Conductor Cross-Sections

The exact measurement of conductor cross-sectional area is crucial to the cable testing.This dissertation proposed a new measurement method based on machine vision detection technology to make up for the demerits of traditional ones.Our machine vision detection system consists of two subsystems: imaging device and image processing system.The functions of each module are described respectively in this dissertation.The imaging device is used for capturing high quality image which constitutes the basis of exact measurement.Components of the device,for example,light-compensating system,optical assemblies,CMOS sensor,are introduced in detail in this dissertation.The experiments were carried out to validate the repeatability accuracy and linearity of our device,by which the imaging coefficient k* was calibrated.The image processing system is composed of two parts: software and hardware.Software is used for the implement of digital image processing algorithm and the interaction with other components.Digital image processing algorithm is the core of the whole machine vision detection system,of which the research contents are as follows:(1)A variety of classical edge detection algorithms are analyzed by comparison,and Otsu method combined with mathematical morphology edge detection is selected as the contour extraction algorithm in this dissertation.(2)According to the characteristics of the noise of cable conductor section,an algorithm based on connected regions labeling is used for reduction of the noise inside the contour of cable conductor.(3)A new method for calculating the global angle of corner detector based on the Chordto-Point Distance Accumulation Technique(CPDA)is applied to perform better in localization of the corners formed by the contour of cable conductor.Based on the advanced CPDA detector,an edge-compensation method is proposed for repairing the edge defects to reconstruct the contour of cable conductor.The software was developed on the basis of the research mentioned above.Efficiency of the machine vision detection system is proved by experiments,which show that the maximum relative error of the measurement method of this research is within 1% and the runtime of each image-process is less than 5 seconds.