Research On Visual Measurement Method Of Geometric Quantity For Part Family Of Disc Brake

With the rapid development of intelligent manufacturing,the quality control requirements for the mechanical parts are getting more and more,and higher requirements are put forward in the measurement efficiency,reliability and accuracy.Non-contact measurement for part geometry can be realized by visual method,which has obvious advantages in automation.However,there are still many problems to be solved in the measurement process planning and measurement reliability for the part family.Based on the analysis of the research status for visual measurement technology,the disc brake part family is taken as the object to carry out the research of geometric visual measurement method.The main work is as follows:Firstly,the overall framework of the visual measurement planning and analysis mechanism has been designed based on the geometric characteristics planning requirements for the disc brake part family.Aiming at the diversity of parts,the strategy of measuring station planning based on 3D virtual model and the strategy of station autonomous matching based on pattern recognition are proposed.After that,a closed-loop process of error chain analysis and behavior optimization was established,and the hardware platform of visual measurement system is built.Secondly,the visual measurement planning and identification sampling strategy for brake disc are studied by starting from the correlation relationship between feature-requirement and planning-collection of part family.Considering the diversity of size series for part family and the limited field of view,the central area and circumferential area are planning separately.For the holes group in the central area,the visual field tolerance relationship,station identification and alignment method are determined.However,for the holes group in the circumferential area,a station identification and alignment strategy based on pattern recognition is proposed.Focusing on the method of acquiring the image samples and extracting feature information,the problem of rapid acquisition of training samples has been solved,and the accuracy of station recognition has been verified.Thirdly,the error sources of visual measurement are analyzed.The influence of operators,such as image gray level transformation,filtering noise reduction,edge extraction and contour fitting on the measurement error of holes group is explored through experimental research.Taking the aperture measurement results as the evaluation basis for the preferred samples,the parameters and the handling process for images are optimized,which effectively reduces the measurement error caused by the image acquisition and image processing process.Finally,a visual measurement and analysis software platform has been built and experimental research has been carried out.The visual measurement and error evaluation mechanism for the roundness and position of holes are established to realize the automatic measurement and analysis for batch parts.The error correction strategy and correction algorithm based on neighborhood search are proposed,and the measurement error is discussed.The experimental results show that the research and optimization of the measurement process behavior is beneficial to improve the reliability and the automation level of the measurement.