Intelligent crack extraction based on terrestrial laser scanning measurement

Abstract

The hazards of cracks, which could badly decrease reliability and safety of structures, are receiving increasing attention with the popularity of tunnel constructions. Traditional crack inspection relies on visual examination, which is time-, cost- and labor-intensive. Therefore, how to identify and measure cracks intelligently is significantly essential. The paper focuses on the Canny method to extract cracks of tunnel structures by the intensity value of reflectivity. We propose and investigate a novel method which combines dilation and the Canny algorithm to identify and extract the cracks automatically and intelligently based on the point cloud data of terrestrial laser scanning measurement. In order for measurement of cracks, the projection of summed edge pixels is adopted, where a synthesis is carried out on the detection results with all sampling parameters. Based on the synthesized image, vertical crack presents two sharp peaks where the space of the peaks indicates the average width of the crack, as well as its position. The advantage of the method is that it does not require determination of Canny detector parameters. The deviation between manual measurement and Canny detection is 2.92%.