Modeling of Crack Depths in Digital Images of Concrete Pavements Using Optical Reflection Properties

摘要

Digital image-based automated pavement crack detection and classification technology has seen vast improvements in the recent years. Although crack lengths and widths can be evaluated using state-of-the-art software with a reasonable accuracy, no reported evidence is found in extending this technology to evaluate crack depths. As a supplement to the existing technology, additional information relevant to pavement crack severity could be revealed by the optical modeling of the image formation process and the subsequent analysis of the variation in pixel intensity profiles within images. A preliminary study was carried out to model the digital image formation of cracked concrete pavements based on the bidirectional reflection distribution function. This study was specifically focused on the optical modeling of shallow longitudinal and transverse cracks as well as joints of concrete pavements using the variation of reflection properties at surface discontinuities. Surface discontinuities were considered to be of regular geometrical shapes for simplification. The new image formation model revealed a definitive relationship among the crack widths and depths and the maximum pixel intensity contrasts seen in the images of the cracks. The model calibration involved the selection of reflection properties to match the pixel intensity contrasts across model generated images of cracks and joints against those of identical cracks formed in concrete pavements. The model predictions of crack depths were also verified using actual crack data not used in the calibration. Finally the usefulness of the calibrated model in evaluating the depths of shallow cracks and differentiating cracks from joints and other surface irregularities in concrete pavements is illustrated.