中文摘要：

为了更精确地评价排水沥青路面的排水特征,采用工业CT对排水沥青混合料马歇尔试件进行扫描,利用数字图像技术获取排水沥青混合料内部结构组成,应用改进大津法进行混合料差异性物质辨识,分割出了空隙图像;对空隙的空间分布特性进行了研究,确定了空隙率、连通空隙率、比表面积与弯曲度的算法;对541幅断面图像遍历处理后获得的空隙数据进行了定量分析,追踪出了排水沥青混合料中的连通空隙路径,并确定了连通空隙路径中的最小截面面积与弯曲度值。研究结果表明：采用改进大津法可有效分割出排水沥青混合料中的空隙图像;利用图像分析结果计算所得连通空隙率与实验室所测连通空隙率基本相同,验证了提出的连通空隙率算法的可靠性;最小截面面积与弯曲度值分别是控制水流流量和排水时间的最重要因素,准确地确定最小截面面积与弯曲度值是精确计算排水沥青路面排水能力的关键;空隙截面面积沿深度方向的分布规律大致为端部空隙面积较大,中间较小,最小截面随机分布;计算的连通空隙率明显小于试件总空隙,可见传统方法中采用总空隙率来评价排水沥青路面的排水特征具有一定的局限性。

英文摘要：

In order to evaluate the seepage characteristics of porous asphalt mixture more precisely, industrial CT was adopted to scan Marshall specimens of porous asphalt mixture. Digital image technology was utilized to obtain the internal structure of porous asphalt mixture. Improved OTSU method was then implemented to identify the different compositions of the asphalt mixture and segment void images. The space distribution characteristics of the voids were also investigated and the algorithms for porosity, connecting porosity, specific surface area and tortuosity were determined. A quantitative analysis of the void data obtained from the traversaltreatment of 541 cross-section images was conducted, therefore the connecting void path inside of the porous asphalt mixture was tracked and the smallest cross-section area and the tortuosity were determined. The results show that improved OTSU method can be used effectively to segment void images from porous asphalt mixture. The connecting porosity obtained by means of image analysis is basically the same with the laboratorial test result, indicating the reliability of connecting porosity algorithm. Since the smallest cross-section area and tortuosity are crucial factors for controls of the water flow and the seepage time, the accurate determinations of these two factors is key to precisely calculate seepage performance of asphalt pavement. Besides, the distribution law of void cross-section in the direction of depth tends to be large at both ends but small in the middle and the smallest cross-section features a random distribution. The fact that the calculated connecting porosity is obviously smaller than the total void of the specimens unveils the limitation of conventional methods in evaluating the seepage of porous asphalt mixture by using total porosity.