AI Solutions for Innovation of Pavement Crack Analysis on Images Taken from Specialized Road Surface Survey Vehicles in Vietnam

In today's world, innovation is not only the application of new technologies, but also the continuous improvement of existing technologies based on new scientific and technological achievements. The article presents the efforts of the authors in researching to improve the technology of road surface survey by specialized road surface survey vehicles, which have been widely used in Vietnam in recent years. On the basis of the road surface image dataset collected during previous surveys, the authors have applied machine learning algorithms to build models to automatically detect pavement cracks on the collected images instead of conventional method of manual detection. Initial efforts using machine learning techniques such as: support vector machines (SVMs), machine learning, especially the combination learning algorithm Boosting (Adaboost) and deep learning (DL) had demonstrated the feasibility of the new solution alternative to manual analysis. The next research at a higher challenging level, a deep architecture using convolutional neural network (CNN) for crack segmentation on gray scale images has been developed for much better improvement of the current technology in terms of pavement crack detection.