Abstract
This paper presents the effectiveness of the error reduction methods for integrated dot centroid tracking method. Two important sources of error and noise are lens distortion and brightness intensity fluctuation on each pixel, which should be taken in account to achieve high accuracy of full-field strain measurement. The method can correct lens distortion at any focal distance between a digital camera and an object. Pixel based noise reduction method using Kalman filter and integrated image by multiple images was proposed. When we need very accurate strain field, the lens distortion should be removed as possible as we can. Due to the corrections, 0.1% strain accuracy is accomplished and total computation time is reduced in comparison with an arithmetic mean operations. The effectiveness of the idea is shown through experimental and numerical examples and the result of strain field measurement is presented and discussed.
