Real-Time Disparity Contrast Combination for Onboard Estimation of the Visibility Distance

Hautiere, N.; Labayrade, R.; Aubert, D. · 2006 · OpenAlex-citations

DOI: 10.1109/tits.2006.874682

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper presents a real-time method for estimating the "mobilized visibility distance" using onboard stereovision cameras, addressing the need for adaptive driving assistance systems in adverse weather conditions. The authors define mobilized visibility distance as the distance to the most distant object on the road surface with a contrast above 5%, a metric closely aligned with the International Commission on Illumination’s definition of meteorological visibility. This estimation allows vehicles to detect fog, adapt sensor operations, or alert drivers when onboard assistance systems become inoperative. The methodology combines two distinct computational tasks: local contrast detection and depth map generation. First, the authors employ an adapted Köhler thresholding technique to identify pixels with local logarithmic contrast exceeding 5%. This approach is chosen for its robustness to Gaussian noise and its alignment with psychophysical contrast definitions. Algorithmic optimizations, such as minimizing the number of thresholds considered and precalculating min-max images, reduce computation time significantly, enabling real-time performance on standard hardware. Second, the system utilizes stereovision to compute depth. Specifically, it employs a "v-disparity" transform, which maps pixel coordinates against disparity values to detect straight lines corresponding to planar surfaces (e.g., the road or vertical obstacles). This allows for the extraction of a robust depth map from the stereo images, filtering out false matches by validating against global surface models. The results demonstrate that the combined method effectively estimates visibility distances in both daytime and nighttime conditions across various meteorological scenarios, including sunny and foggy weather. The system processes images within 60 ms on a standard computer, meeting real-time requirements for onboard implementation. The estimated mobilized visibility distance correlates well with meteorological visibility standards, providing a reliable measure of the effective range of exteroceptive sensors. The significance of this work lies in its ability to provide a generic, real-time visibility estimation tool that does not rely on fixed infrastructure or specific road markings. By integrating contrast analysis with stereoscopic depth perception, the method offers a practical solution for enhancing the reliability and safety of autonomous driving systems and driver assistance technologies in low-visibility environments. This approach enables vehicles to dynamically adjust their operational parameters based on actual atmospheric conditions, thereby improving overall system robustness.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• visibility analysis litigation
• weather rain fog snow
• disability glare