2019 CSCE Annual Conference - Laval (Greater Montreal) Conference
Dr. Naixing Liang
Dr. Susan Tighe, CPATT - University of Waterloo
Focus on the problem of continuous real-time monitoring of aggregate distribution uniformity during asphalt pavement paving process, an asphalt mixture separation and identification method based on image processing technology and a uniformity evaluation method based on static moment theory were proposed. Firstly, the colour photos shot from 60cm high was preprocessed by graying, filtering and equalization to get the high contrast and low noise gray image. And then, using the image threshold segmentation method and morphological image processing to separate the aggregate and gain the binary image. In this process, we develop the watershed segmentation method with the Extended-Maxima transform, which can significantly eliminate the over-segmentation of the image, and achieve fast and accurate segmentation of the adhesion particles. Considering the influence of aggregate size and distribution position on the uniformity, the product of aggregate particle area and distance to image boundary is used as the basis for uniformity evaluation. The concept of uniformity is proposed, and then the boundary criteria were determined by analyzing 1,000 random on-site collected paving photos. When the uniformity is less than 0.91, the aggregate is in the uneven state, and vice versa. Moreover, the remaining 85 photos verified the accuracy of the demarcation criteria. The example and validation test indicate that the uniformity analysis method may be efficiently applied to the evaluation of aggregate distribution on asphalt mixture during the paving process.