A Vision-based System for Extension Measurement
TIn material testing, video measurement has the following advantages: (1) lack of influence on the test specimen, (2) decreased problems with knife-edge slippage, etc. In this paper, a machine vision-based extension measurement system using image processing techniques was developed. By attaching two stickers on the specimen under tensile testing, a sub-pixel image processing scheme was developed to measure the distance between two stickers automatically. The influence of image interpolation, sticker size, and field of view on the accuracy of the measurement was investigated. Testing of the system on a rubber specimen was done. Inverse distorted calibration and bilinear shape function calibration methods were considered and compared. The experiment results show that an error of 1.5% can be achieved with a proper size of sticker, smaller field of view, and appropriate calibration method.
Keywords – Image processing, Sticker, Sub-pixel, Extension measurement.