A set of technical solutions for fast acquisition of bridge crack images, automatic information recognition and apparent panoramic reconstruction is proposed in this paper. High-quality, large-sized and fast acquisition of crack images is achieved by virtue of the study on the number of cameras, shooting parameters, shooting height, image coding rules and feature-based stitching algorithms. Besides, through the study on image preprocessing technique and crack recognition algorithm, the automatic identification of crack information is realized. What’s more, a program for automatic crack recognition and panoramic 3D reconstruction was developed by means of OpenCV+OSG. The research results are applied to the intelligent bridge detection system, which is helpful to improve the accuracy of crack identification and automation and information level of bridge detection.

CICTP