主催: 一般社団法人 日本機械学会
会議名: 第11回マイクロ・ナノ工学シンポジウム
開催日: 2020/10/26 - 2020/10/28
Aluminum alloy sheets are sometimes cracked during bending process. Currently, the degree of cracking is evaluated visually, but there are some problems such as cost and evaluation variability. Therefore, a fast and quantitative evaluation method for bending cracks is required. In this study, we attempted to quantify bending cracks using image processing. 1) The tendency of the crack occurrence was grasped from the observation of the aluminum sheet sample cross section. 2) The optimal imaging conditions for crack detection were determined by optical considerations. 3) The bending cracks were quantified as the number of pixels using image processing. Finally, the calculated bending crack scores showed a high conformance to the visual evaluation.