Research On On-line Detection Technology Of Machine Vision For Commutator Inner Hole Defect

The commutator is one of the core components of the motor,and its demand is increasing with the development of the motor industry.At present,the inner hole quality inspection of the commutator mostly uses the manual observation method,and the defect cannot be quantitatively described,and the inner hole is small,the detection process is subjective,and the detection process is inefficient and inefficient.In order to improve the detection efficiency of the product and adapt to the development of intelligent manufacturing,this paper proposes a set of machine vision automatic detection scheme for internal hole defects.In this paper,the characteristics of the inner hole and the defect are analyzed,and the overall scheme design of the detection system is carried out,including the selection of the optical device,the design of the illumination mode,the design of the image acquisition scheme and the division of the software module.The calibration method of the detection system was studied.The distortion type of the inner hole detection lens was analyzed.The influence of camera position and deviation on the imaging results was explored.The calibration target and calibration tool were designed.The calibration method of the detection system was determined and the calibration experiment was carried out.And calibration accuracy analysis.An image expansion algorithm is designed for the acquired non-equal resolution inner hole image,which realizes the equal resolution expansion of the inner ring image.Aiming at the problem of uneven illumination and the incomplete detection of the inner hole detection area in the unfolded image,a de-lighting uneven algorithm based on Lab color space and a sharpness enhancement algorithm based on image fusion are proposed to realize the unfolded image.The quality is enhanced.An internal hole damage inclusion defect detection algorithm combining gray-scale morphology and dynamic threshold is designed for common types of defects in the inner hole,and a defect detection algorithm based on color difference component is proposed.The experimental verification of the internal hole detection technology is carried out,including the robustness verification of the expansion algorithm,the adaptive verification of the image enhancement algorithm,the accuracy verification of the defect detection algorithm and the time efficiency analysis of the detection.In this paper,the quality of the inner hole of the commutator with a diameter of 9.84 mm and a depth of 10 mm is realized in the built-in inner hole testing and development experimental bench.The detection accuracy is high,and the speed is fast,which can realize the requirement of subsequent inner hole online detection.