Recently, computer-assisted systems for reliable detection of surface cracking of the concrete have been developed as an alternative for handcrafted methods [2], [3]. Most of the computer systems are based on the digital image analysis, which can be performed semi-automatically [4]. In order to detect cracks, users of the systems need to specify the focal part of the element for inspection. However, the cracks cannot be detected accurately when the images include noise. In this case, a manual interaction must be done. Consequently, such systems cannot be used in the fully-automated analysis. In attempting to realize an automated crack detection system, a more reliable digital image processing technique is essential. In the engineering practice, Digital Image Correlation (DIC) technique has been implemented for reliable detection of concrete surface cracking, e.g. [5], [6], and [7].