基于机器视觉的电机换向器质量在线检测系统开发

电机换向器质量检测是换向器生产线中的一个重要工序,但其仍采用人工方式,存在检测效率低、漏检率高等问题。针对此问题,本文运用图像处理和机器视觉技术,开发换向器质量在线视觉检测系统。该系统针对生产过程工艺多变,造成形状检测标准不一致问题,提出自适应学习模板方法；在轴孔孔径检测,提出基于Freeman链码改进的孔径快速检测算法；在端面缺陷中,提出基于改进视觉注意力模型的端面缺陷检测方法。实验结果表明,系统检测精度达到99.80%,漏检率为0%,F-measure值为99.89%。该系统能够快速有效检测换向器存在的外观质量问题,可满足换向器在线质量检测需求。

A Development of On-line Quality Defection System for Electric Machinery Commutator Using Machine Vision

A quality detection is an important process in the production line of electric machinery commutator. The inspecting mode still dependent on human, leads low speed of detection and low accuracy. A system for on-line inspection quality of electric machinery commutator is developed using a machine vision technology. Aiming the problems which there exists different inspecting standards of commutator shape because of the production process, an approach of adaptive learning templet to detect the shape is proposed. An improved method to detect the diameter of axle hole is based on Freeman chain code. Aimming to defect flaw of end-surface of commutator, an approach based on a vision-attention model is proposed. Experimental results show that inspection accuracy of the proposed system reaches 99.27%, miss rate of 0%, and the F - measure value of 99.59%. The system can quickly and effectively detect appearance quality of the commutator, which can meet the needs of the commutator line quality inspection .