关键词: 桥梁工程, 桥梁裂缝, 自动检测, 立体视觉

Abstract: Bridge crack is an important indicator for bridge safety evaluation. Aiming at the disadvantages of high error, high missing detection rate and poor real-time capability in the traditional detection method of bridge crack, the bridge crack automatic detection system based on stereo vision was designed, combining with the digital image processing technology. The system obtained the feature points’ coordinates through camera calibration, setting feature points and preprocessing the crack images, and then the binocular distance measurement principle was applied to measure the crack depth. Moreover, the length and width of the crack was calculated by the statistics of related binary image pixels. The experimental results show that the proposed system can accurately measure bridge crack’s length, width and depth online in real-time.

Key words: bridge engineering, bridge crack, automatic detection, stereo vision