Tack Coat Image Analysis using a Combination of Mask R-CNN and GLCM

Tack Coat is a thin layer of asphalt between the existing pavement and asphalt overlay which contributes significantly to the long-term durability of the asphalt pavement layers. When the tack coat materials are being applied to the pavement, it is important to observe the uniformity of the tack coat application. Conventional evaluation approaches are expensive and time-consuming. Furthermore, effective data collection methods are also desired by the road monitoring agencies. Recently intelligent road monitoring has become popular. Different automated methods such as machine learning, deep learning and computer vision-based systems, have been developed in this area. But the majority of the existing systems are developed for paved roads and limited bounding boxes for feature extraction and classification. In this work, we are proposing a deep learning framework using the Mask R-CNN to do instance segmentation on the image to determine the tack coat region and apply GLCM to perform the texture analysis to measure the uniformity and non-uniformity coverage of the tack coat on pavements. We tested our model with other state-of-the-art algorithms and it showed promising performance on the segmentation and classification.