Research On Multi-parameter Detection System For High Precision And Small Parts

The measurement of precision parts in the manufacturing industry has always been regarded as a basic task.The measurement of the size of domestic parts,especially small parts,is usually done manually,not only time-consuming and laborious,but also the measurement accuracy is low.Therefore,the measurement efficiency and measurement accuracy of small parts need to be improved.With the increasing maturity of machine vision technology in the measurement field,it brings new ideas to solve this problem.In this paper,a multi-parameter detection system based on machine vision is designed based on the key groove parts.Through image processing technology,precision parts are measured with high precision,multi-parameter and non-contact.The main tasks of the project are as follows:(1)The visual detection system of precision parts is designed.According to the testing requirements,complete the hardware part design and related software design.The hardware part includes the selection of cameras,lenses,and light sources.The software includes the design of some column image processing algorithms.(2)The image preprocessing technology is studied.The image gray transformation,image de-noising,and image segmentation algorithms are studied.After experimental verification and analysis,the median filtering method is selected to remove noise,and a fixed threshold is selected to segment the part image and extract the parts of interest.To facilitate the next step in image edge detection.(3)The image edge detection technique is studied,the classical edge detection algorithm is analyzed,and the edge of the part is extracted by the Canny algorithm.Based on this,this paper analyzes and studies the sub-pixel detection algorithm of the part image,selects the bilinear interpolation algorithm to detect the sub-pixel of the part,and uses the least squares method to fit the edge outline to complete the part parameter detection and improve the detection accuracy.(4)A visual inspection experiment platform is set up to calibrate the detection system,complete the conversion between the pixel size and the actual physical size,and detect the actual size parameters of the parts.On the VS2015 platform,Halcon combined with C# to complete the system's human-computer interaction interface design and display the measurement results of component parameters in real time.The experimental results show that the accuracy of the system is within ± 6?m,which verifies the feasibility of using visual detection method to measure the parameters of high-precision parts.To sum up,through the visual detection system design,image processing algorithm analysis research,experimental platform construction,multi-parameter measurement experiments,this paper verifies the correctness of the visual detection scheme,and lays a solid foundation for the follow-up high-precision,non-contact measurement.