Strain measuring technology based on machine vision

Aimed at the shortage of traditional contract extensometer measuring method on the tensile test of metal material, the principle of strain measurement based on machine vision was applied to specimen with non-contact precision strain measurement in this paper. To the image preprocessed, a straight line was got by probabilistic Hough transform. Then the edge was refined by using the sub-pixel, finally, a new line was got as the target line combined with the least square linear fitting method. Afterwards, we calculated the pixel length of the target line to get the actual size of the specimen. The test results show that the method is effective, which provides the foundation to the research of high precision strain measurement technique.