Video-based Detection and Classification of US Traffic Signs and Mile Markers using Color Candidate Extraction and Feature-based Recognition

Traffic sign and mile-marker detection and classification are among the important components of highway asset management systems. The significant number of these high-quantity and low-cost assets in US highways can negatively impact on quality of any manual data collection and analysis. To address these challenges, this paper presents an efficient pipeline for video-based detection and classification of traffic signs and mile-markers based on color and shape criteria. Candidate extraction is based on finding the optimum RGB thresholds which yields high detection rate (very low False-Negatives), while keeping the number of FalsePositives in check. The connected components from a thresholded image are extracted next. We use sliding windows, Haar-like features, and AdaBoost learning method to classify the detected assets. Experimental results with an average classification accuracy of 79.30% on actual data collected from US-460 highway show the promise of the proposed method for reducing the time and effort required for developing traffic road asset inventories.