Lane Departure System Design using with IR Camera for Night-time Road Conditions

Miman, Mehmet; Akırmak, Osman Onur; Korkmaz, Hakkı Can · 2015 · OpenAlex-citations

DOI: 10.18421/tem41-06

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 addresses the critical safety issue of lane departure detection during nighttime driving, a period associated with a disproportionate number of traffic accidents. While existing lane detection algorithms perform well in daylight, they often fail under low-light conditions. The authors propose a vision-based system designed to detect when a vehicle unintentionally leaves its lane at night, issuing warnings via a Head-Up Display (HUD) to alert the driver. The primary motivation is to enhance road safety by leveraging infrared (IR) camera technology, which offers superior visibility compared to standard RGB cameras in dark environments. The study employs a MATLAB-based algorithmic approach using images captured by a single forward-facing IR camera mounted on the vehicle’s rearview mirror. The processing pipeline begins by converting RGB images to grayscale and applying background subtraction to isolate road markings. Edge detection is then performed to preserve lane marking features, followed by the Hough Transform to identify straight lines in parameter space. To improve accuracy, the algorithm restricts the search angle to -60 to 60 degrees and filters out noise by removing pixel groupings smaller than 15 pixels. For video streams, the system tracks lane markers across frames, validating their presence over multiple frames to handle occlusions or low-quality video. The system identifies left and right lane boundaries and triggers a HUD warning if the vehicle crosses these markers. Experimental results demonstrate that the proposed algorithm effectively detects lanes in nighttime conditions where standard daylight cameras failed. The IR camera enhanced the visual contrast of white lane markings, allowing the Hough Transform to accurately identify long straight lines corresponding to lane boundaries. The authors note that while the algorithm performs with good precision and robustness in controlled nighttime scenarios, it is susceptible to noise from complex road conditions. The system successfully matched detected lines with real-world lane markers in test cases, proving the viability of using a single IR camera for this application. The significance of this work lies in demonstrating that cost-effective, single-camera IR systems can provide reliable lane departure warnings at night, potentially reducing accidents caused by driver fatigue or inattention. The authors conclude that while the current algorithm is effective, future improvements should include multi-sensor fusion and advanced models like B-Snake to handle complex urban environments, such as curved roads, shadows, and irregular markings. Implementation on hardware like FPGAs or DSPs is recommended for practical automotive deployment. This research contributes to the field of driver assistance systems by validating IR-based vision processing as a viable solution for nighttime safety applications.

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