Detection and tracking of traffic signs using a recursive Bayesian decision framework

Abstract

Abstract-In this paper we propose a new method for the automatic detection and tracking of road traffic signs using an on-board single camera. This method aims to increase the reliability of the detections such that it can boost the performance of any traffic sign recognition scheme. The proposed approach exploits a combination of different features, such as color, appearance, and tracking information. This information is introduced into a recursive Bayesian decision framework, in which prior probabilities are dynamically adapted to tracking results. This decision scheme obtains a number of candidate regions in the image, according to their HS (Hue-Saturation). Finally, a Kalman filter with an adaptive noise tuning provides the required time and spatial coherence to the estimates. Results have shown that the proposed method achieves high detection rates in challenging scenarios, including illumination changes, rapid motion and significant perspective distortion.

Efficient hybrid monocular-stereo approach to on-board video-based traffic sign detection and tracking

“…Namely, in 3D the linear process modeled by the Kalman filter corresponds to an inherently linear process in the 3D parametersin contrast to the 2D tracking case. The enhancement is exemplified in Figures 16 and 17, which shows tracking of the same traffic sign in 2D (using the algorithm in [7]) and 3D, respectively. As can be observed in Figure 16, as the TS gets closer the sign, the bounding box delivered by the 2D tracker loses accuracy, while the 3D tracker perfectly fits the sign bounding box.…”

“…All candidates not fulfilling these constraints are disregarded in order to avoid unnecessary processing overhead, especially when it comes to 3D reconstruction. Details on the functions modeling the TS area, pictogram area and aspect ratio can be found in [7].…”

“…From: http://proceedings.spiedigitallibrary.org/ on 05/14/2015 Terms of Use: http://spiedl.org/termsFigure 16: tracking example of a given sequence using the monocular approach in[7].…”

Traffic Sign Detection and Tracking Using Robust 3D Analysis

“…A more detailed explanation can be found in [7]. First, the color segmentation stage aims to separate the TSs from the background using the color information of the objects according to the Hue (H) and Saturation (S) components of HSV color space.…”

Extraction of road traffic sign information based on a vehicle‐borne mobile photogrammetric system

“… Carry out image edge detection by highlighting the contours (outlines) of traffic signs first and then detect true traffic signs through shape analysis of the contours (Alefs et al., ; Jiménez et al., ; Deguchi et al., ; Gu et al., ; Zavadil et al., ). First, conduct image segmentation with red/green/blue (RGB), hue/saturation/intensity (HSI) or other colour spaces for identifying candidate regions of traffic signs; then confirm the true traffic signs through the shape analysis of candidates (Gao et al., ; Prieto and Allen, ; Xu, ; Marinas et al., ; Song and Liu, ; Zhang et al., ). …”