Electric Machinery Commutator Appearance Many Defects Of On-line Inspection Based On Machine Vision

With the rapid development of China's economic construction, the motor as defense, transportation, industrial and daily life indispensable foundation device, wherein the motor commutator one of the core components, but also with the development of the motor industry and gradually increase Commonly known as commutator, it is an important part of the continuing operation of the motor, and its type is divided into:hook type and groove type. Wherein the appearance of quality testing of the commutator production line is an important step with the growing demand for industrial automation, commutator industry is also in urgent need of upgrading technology, greatly improve the production line of intelligent, level of automation has become an inevitable trend.Since the commutator during production process due to various factors, material quality problems caused by the presence of multiple commutator, characterized by different defects, mainly in the shape defects and surface defects. Specific bad shape defect can be divided into hook (off hook, partial hook, stick hook), the presence of foreign body hook groove, aperture unqualified; commutator surface defects mainly refers to the presence of upper and lower end Bakelite defect exists cracks, the surface of copper powder; faulty product which is to be removed before the sale of the commutator, the commutator appearance quality detection line in the commutator is an important step. At present, the quality inspection of the commutator still use manual inspection manner, depending on the person testing experience and subjective judgments, the low efficiency of its detection, false positive rate and high rate of missed, it is difficult to provide a reliable, accurate and stable Test results. In addition, the detection process accounted for about 25% of the total workforce of the production line, resulting in increased production costs; and long-term work on the human eye and also has a larger cervical spine injury. These factors have affected the commutator manufacturer's production quality and efficiency.For detecting the presence of said commutator quality problems, paper for the detection of objects commutator, commutator develop quality online detection system based on machine vision, performs the following tasks:(1) Against the commutator manufacturing process variability, leading to inconsistent standards and shape detecting the presence of a high false alarm rate and low adaptability problem, adaptive learning method, by learning about the different types of products commutator corresponding test data, learning to use the data obtained as a detection standards for product testing.(2) For the shaft hole diameter manual measurement precision and speed are insufficient presence, proposes the use of efficient detection based on Freeman chain code to improve the symmetry of the four-way shaft hole algorithm, by extracting the key pixels reduces the computational load while maintaining measurement the accuracy of the test results show that the design can meet the production requirements.(3) Classification for the proposed defect detection algorithm based on multi-feature SVM in the end surface defect detection, the algorithm in the source image based on image preprocessing, extracting a plurality of feature image mean, variance, entropy and other energy feature vectors through simulation and actual test SVM classifier, in the case of small sample selection of SVM classifiers can achieve better classification results, the accuracy rate of more than 94.59%.(4) Finally, in the Win 7 platform, based on Mircrosoft Visual Studio2010 development of quality on-line visual inspection system for commutator and debug. Simulation results show that the development of a visual inspection system can quickly and accurately detect the presence of various defects commutator, meet the design requirements.