Tensile Strain Measurement Of Materials Based On Machine Vision

In this paper,a method of measuring the strain of material sample with low resolution CCD camera is proposed,which can meet the requirements of material mechanical property test.This method corrects the lens distortion in a single direction,simplifies the solution of the lens distortion mathematical model,establishes the measurement model of the imaging plane and the measurement plane which are not coplanar,and uses a parameter to express the influence of the object distance change on the measurement result.The experimental results show that the method proposed in this paper has the advantages of high accuracy and good uniformity.?1?In this paper,six sub-pixel edge detection algorithms in one-dimensional case are analyzed,and the evaluation method of block target detection is proposed,Under the evaluation standard of detection accuracy and detection speed,the gauge block target is measured.The results show that the one-dimensional gray-scale moment method has high accuracy and fast detection speed,and is suitable for strain measurement of material specimens.?2?According to the characteristics of one-dimensional deformation of materials in tension and compression,a method of correcting lens distortion along the deformation direction is proposed in this paper.The influence of the window size and distortion correction of the homography matrix on the deformation measurement accuracy is analyzed by experiment and calculation.The measurement window of 190×190 pixels and the distortion correction of 4^th-degree polynomial can meet the requirements of online measurement speed and accuracy.?3?The measurement model of the non coplanarity between the measured plane and the imaging plane is established.The results show that the maximum difference between the calculated strain of this method and that of the extensometer is 0.35×10^-4 and 0.41×10^-4,respectively,accounting for 0.63%and 1.07%of the total strain,which shows that this method has good accuracy.