High-accuracy crack detection for concrete bridge based on sub-pixel

Crack is one of the main reasons for the damage of Concrete Bridges. Crack detection is a critical responsibility and essential task. The accuracy of crack detection is very important. A high-accuracy edge detection algorithm at sub-pixel level is proposed in this work. A blurred edge model is adopted here, and a least-squared-error based solution is derived. Compared with other algorithms, the algorithm proposed in this paper creatively uses the Sigmoid model, this makes the fitting error smaller and the solving process simpler. Its applications to both synthetic and real images are presented for evaluation. Finally, the algorithm is applied to crack detection of concrete Bridges based on climbing robot, and a high-resolution camera is used to collect crack images. It is very difficult to distinguish real fractures from false ones without damaging the bridge. We use the red laser line to illuminate the crack, then we detect the breakpoints of the laser line, by this way, the false crack and noise interference can be.