Stereovision-based surface flaws detection experiment of foundations of Yiqiao Bridge

Abstract

To evaluate surface flaws on the foundations of bridge structures, a stereovision-based surface flaws detection method is proposed to evaluate the geometry size of surface flaws. Both scale invariant feature transform algorithm and a feature extraction algorithm based on area matching are adopted to perform feature extraction and matching on the left and right flaw images. The three-dimensional coordinates of feature point are recovered based on the parallel stereovision model; the length and depth of surface flaw can be evaluated using the corresponding mathematics formula. Besides, two area calculation methods are employed to compute the area of surface flaw in this article. One method is accumulation of area of each triangle which is obtained by dividing surface flaw; another method is to adopt the Monte Carlo method to calculate flaw area. In situ test of surface flaws is conducted on the foundations of Yiqiao Bridge. The surface flaws such as spalling, crack, and hole are measured by the stereovision measurement system. Experimental results indicate that the stereovision-based surface flaw detection method can accurately measure the geometry size of surface flaws; this verifies that the proposed method is reliable and useful for evaluating surface flaws on structures.