An image processing system for health monitoring of road pavement

本研究提出一种用于路面健康监测的图像处理系统。作为第一步，图像滤波器从各种路面图像中准确地检测出裂缝。然后，为了有效地利用滤波器，我们提出了两种方法作为应用滤波器的预处理。一种是利用最大似然分类器对路面类型（混凝土或沥青）进行分类的方法.另一种方法是从图像中识别出汽车可能主要经过的区域。最后，这些方法结合成一个系统，它使我们能够自动评估健康水平的路面从图像和指示点，需要维护。

This research presents an image processing system for health monitoring of roadpavements. As the first step, an image filter to detect cracks accurately from variouskinds of images of road pavement. Then, in order to utilize the filter effectively, wepropose two methods as pre-processes for applying the filter. One is a method to classifytype of pavement (concrete or asphalt) by using maximum likelihood classifier. Anotheris a method to recognize the region where cars might mainly pass from the images. Finally,these methods are combined into one system and it enables us to automatically evaluatehealth level of the pavement from the images and indicate points where maintenance is required.