Size Measurement System For Fuel Pump Housing Based On Machine Vision

With the rapid development of modern industry,the requirements of measurement technology are also getting higher and higher.Many non-contact measurement methods come into being when the traditional measurement methods encounter the bottleneck of development Among them,the machine vision measurement technology has become the focus of researchand application in the field of industrial detection.This technology uses industrial cameras to simulate the human visual function,and takes the workpiece image captured by the camera as the input of the measurement system,then processes the workpiece image via computer to obtain the geometric structure information of the workpiece,finally output the measurement results of the workpiece.In order to conform the current development trend of automobile parts detection,this dissertation researches and develops a vision measurement system for automobile fuel pump housing on the basis of studying the development status of machine vision at home and abroad.The main contents of this dissertation include system hardware architecture design,improved supixel edge detection,subpixel-based size measurement for fuel pump housing,system software design and measurement experiments.Firstly,this dissertation designs the corresponding light source and lighting schemes according to the two measurement indexes of the fuel pump housing,and a complete vision measurement system is built combined with other hardware.Then,this dissertation proposes an improved sub-pixel edge detection algorithm based on quadratic curve fitting after full research on classic pixel-level edge detection algorithm and sub-pixel edge detection algorithm,and verifies that the performance of the improved algorithm is better than the original algorithm through experiment After that,the camera calibration method of this system is researched,and a subpixel-based size measurement method for fuel pump housing is proposed according to the analysis of imaging of two measurement indexes,and elaborates the least-squares curve fitting algorithm involved.Finally,we complete the software design and test the measurement accuracy of this system through experiment.The experimental results show that this vision measurement system has high repeated accuracy and high measurement speed,which can meet the accuracy and real-time requirements of industrial measuremen.