Host: The Japan Society of Mechanical Engineers
Name : [in Japanese]
Date : June 28, 2023 - July 01, 2023
Social infrastructure is in urgent need of inspection and maintenance due to aging problems. In order to solve this problem, research and development of inspection robots is underway. The images captured by the camera mounted on the inspection robot include blurred images. Visual inspection of concrete measures crack widths of 0.05 mm or more, so if the image taken by the inspection robot contains blurring, the crack width cannot be measured accurately. Therefore, it is necessary to develop a system that automatically distinguishes between blurred and non-blurred images. In this paper, we describe the discrimination evaluation by quantitatively representing the blur of the image taken by the inspection robot. We evaluate the proposed method and reveal suitable features by discriminating blurred and unblurred images using a crack detection system.
