Research On Metal Rod Parameter Measurement System Based On Machine Vision

With the development of computer technology,machine vision is widely used in dimensional measurement in the industrial field.Compared with the traditional manual measurement method,the visual measurement based on Charge-Coupled Device(CCD)has the advantages of high precision,non-contact and high reliability.With the gradual improvement of the social automation level,there is an urgent need for equipment that can automatically measure the production parameters of the workpiece.Therefore,this paper mainly studies the CCD sensor measurement system based on the single-chip microcomputer control,which can realize the measurement of the straightness,diameter and roundness of the metal rod.The STM32F103 controller is used to drive the linear array CCD in time sequence,and the output signal of the CCD sensor is filtered and amplified,A/D conversion and other corresponding circuits are processed,and then the processed signal is transmitted to the host computer through the serial communication,and then the collected one-dimensional image data is processed by the host computer to realize the measurement of the straightness of the metal rod.The STM32F407 controller is used to drive the area array CCD in time sequence.The output signal of the CCD sensor is filtered and amplified,A/D conversion and data storage are processed by the corresponding circuit,and then the processed signal is transmitted to the host computer through serial communication.The upper computer processes the collected two-dimensional image data to realize the measurement of the diameter and roundness of the metal rod.By comparing different filtering algorithms,this paper adopts wavelet transform with multi-scale characteristics,and by comparing the advantages and disadvantages of soft and hard threshold functions,a wavelet algorithm with improved threshold function is proposed.Experiments show that the noise reduction effect of this algorithm is better than soft and hard threshold functions.In order to separate the target and the background,this paper compares the global threshold and local threshold segmentation methods,and finally adopts the global threshold segmentation method based on the genetic algorithm to optimize the threshold to complete the segmentation of the target and background.The Sobel operator is used to extract the edge feature of the target.Considering that the Sobel operator in two directions cannot accurately locate the edge,so the Sobel operator is increased to eight directions,and the convolution template is improved accordingly.By comparing different sub-pixel methods,the interpolation method is selected to interpolate the pixels,sub-pixel detection is performed on the target edge,and the sub-pixel edge features are extracted.Finally,the measurement of the straightness and roundness of the metal rod is realized by the evaluation method based on the minimum area method.