Pavement Disease Detection Method Based On Image Analysis And System Development,

Pavement distress detection plays an important role in maintaining pavement. The expenses of the pavement maintenance can be reduced greatly if the distress is detected and administrated in the initial stage. The traditional detection methods based on artificial vision can not meet the need of rapid development of the road for its low-efficiency, non-precision, inconvenience and etc. Therefore, the approaches of pavement distress detection and system development based on images have drawn the world attention.The detection of the crack distress has been focused on in this paper. The following threshold segmentation will come across various difficulties for the uneven illumination of the pavement imaging. Therefore, it is necessary to conduct the gray correction in imaging pretreatment stage. By analyzing the signal model of pavement image, the image's background sub sets can directly be obtained from original image. Then based on the interpolation technology, the pavement's background image can be gotten immediately. Finally, through adopting the additive model to distract the background image from the original one, the revised result can be obtained naturally.After the gray correction, the threshold segmentation can be calculated with gray histogram, the two-valued processing for distressed image can be conducted and the distress information can be drawn, due to the accounting feature of pavement distress showed in the image. Because of the obvious borderline of the crack distress, the traditional measurement can reinforce the target information. Eventually, the complete line-form distress information can be achieved after removing the noise and non-crack target.For the massive crack, the cut-off crack can be connected by enlarging the connecting area based on the contour-tracing, because the crack, detected by threshold segmentation, is not continuous. Such measurement is beneficial for the experiment on the massive distress images.A series of off-line data process system of pavement distress have been introduced in this paper, which are applicable for the storage and measurement of the massive distress images. Through experiments, this system can highly improve the efficiency of image processing. Meanwhile, it can do preparation for the establishment of the perfect database for distress image.