Video-Based Highway Asset Recognition and 3D Localization

A key element towards development of an asset management program is an efficient data collection on high-quantity, low-cost highway assets. Despite the importance, current practices of asset data collection are still manual and time-consuming. There is a need for a well-managed asset data collection that can provide usable asset inventories to Departments of Transportation (DOTs) for further analysis and condition assessment purposes. In this paper, we present a novel video-based recognition and 3D reconstruction algorithm. Our method takes an input of video streams and combines 2D recognition with 3D reconstruction algorithms. Using a new Support Vector Machine (SVM) classifier and based on the color channels at pixel level, a set of bounding boxes is initially extracted. Using a Haar-based shape recognition algorithm, the 2D candidates are further categorized based on their shape. These candidates are placed into a texture+color recognition algorithm. The benefits and limitations of the method in detection, classification, and localization of multiple types of assets are discussed in detail.