New Crack‐Imaging Procedure Using Spatial Autocorrelation Function

Publisher Website
Google Scholar
Abstract
An innovative method of applying a spatial autocorrelation function to analyze pavement image data is presented in this paper. One of the major advantages of using a spatial autocorrelation function is its capability of suppressing the noise. By calculating the autocorrelation function of the pavement image, we can measure direction and area of the crack under the presence of background noise. An algorithm is developed and programmed using C language to identify crack type and to measure crack density automatically. The algorithm is illustrated using artificially prepared crack images. Randomly selected sample video images of real cracks are used to demonstrate the accuracy of the proposed automated crack‐imaging procedure. Based on the limited set of data, it is concluded that the proposed automated procedure can identify crack type and density with a reasonable accuracy. It is proposed that the crack‐density measure could be used as a combined index of various crack types for managing pavements at the network level.

This publication has 6 references indexed in Scilit: