Research On Image Detection Method For Appearance Defects Of Automotive Air Conditioning Compressors

Nowadays,the Chinese society is developing rapidly,people’s living standards are getting higher and higher,the number of cars is increasing,and the air-conditioning system has a great influence on the comfort of automobiles.It has become an indispensable and important component system for automobiles.For the problem of obtaining the plane distance for the circular holes of different planes of the air-conditioner compressor,the traditional detection methods have many problems such as low precision and slow detection speed,which cannot meet the requirements of detection and capacity.The development of machine vision and digital image processing technology has made it possible to automate the judgment of product dimensions.In order to solve the problem of plane distance for the circular holes of different planes of air-conditioner compressor,a set of on-line detection system with automatic support for measuring the plane distance of the hole including hardware support,algorithm analysis and software execution is designed.Specific research content includes:Air compressor compressor round hole image acquisition system design.The system has improved the traditional coordinate measuring machine,and uses 3D sensor and improved measuring device to make the measurement.Through the position adjustment of the 3D sensor and the mechanical structure,the relevant parts of the size to be measured are scanned and sampled,and then the sampling is performed.The figure performs image processing.The measurement principle of the selected 3D smart sensor was analyzed and the model of the 3D sensor was determined.In this system,the 3D smart sensor can only obtain the(x,z)two-dimensional coordinates of the air conditioner compressor,and combine the displacement and rotation of the servo system to obtain the complete three-dimensional coordinates of the surface of the air conditioner compressor.Air conditioning compressor round hole extraction.Firstly,the point cloud data is preprocessed,and the point cloud denoising and point cloud segmentation techniques are mainly studied.In the point cloud segmentation stage,an improved RANSAC algorithm is proposed.The effects of RANSAC algorithm and improved RANSAC algorithm are compared,and the accuracy of segmentation is improved.When performing point cloud segmentation,a reasonable judgment threshold,a distance threshold,and an angle threshold are set in advance,and the distance between the sampling point and the initial model plane and the angle of the angle are controlled to ensure the integrity of the surface data of the point cloud to be segmented,so as to avoid Over-segmentation results in optimized segmentation results.Then,the effects of Roberts,Sobel,Laplacian and Canny operators on the extraction of point cloud boundaries are compared.The Canny algorithm is the best to extract point cloud boundaries.Finally,the center of the circular hole is fitted by least squares.Coordinate transformation and projection calculation.The three-dimensional model is transformed into a two-dimensional image by model transformation,perspective transformation,projection transformation,and window transformation.The three-dimensional coordinates of the air conditioner compressor are converted into two-dimensional coordinates by coordinate transformation,and the measurement of the plane distance of the circular holes in different planes of the air conditioner compressor is realized.System function integration and interface design.The paper uses Visual Studio 2015’s MFC framework for software design,and the algorithm function calls the PCL point cloud processing library for point cloud file processing.Finally,the point cloud data two-dimensional calculation operation is realized by calling the OpenGL interface.