Pavement crack detection method based on multi-image and multi-resolution

The complexity of the pavement image and the feebleness of crack signal in pavement make its crack detection very difficult.Therefore,proposed a road surface crack detection methods was proposed based on multi-image and multi-resolution treatment.As for the data acquisition,two cameras with different visual angles on the same road segment were used for data acquisition and illumination with directional light and natural light was used to carry out the test respectively.The image pyramid transform was used for multi-scale decomposition of the image,the threshold of the decomposed image in every scale was processed,and processing on each decomposition scale of image for reconstruction a gradient-like image was reconstructed.Then the watershed algorithm was used to segment the gradient-like image to get a fine image for crack detection.Finally,the two detection images of the same road segment were fused to get the final detection image.The experiment proved that the detection effect of the fused image was better than single image,the use of multi-resolution method could reserve more geometric characteristics of images,and the reliability and precision of the crack detection were improved to a large extent.