Millimeter Wave Radar Guided Yolov8-Based Camera Perception for Vehicle Detection in Indian Traffic
Computer and Decision Making An International Journal · 30.10.2026
DOI: 10.1922/comdem.v3i6s.18781 ↗Kurz gefasst
Fachartikel in der Zeitschrift Computer and Decision Making An International Journal, online erschienen am 30.10.2026. Verfasst von Udhaya Kumar und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
The reliable perception of vehicles in a heterogeneous traffic scenario is a key requirement for intelligent transportation systems and Advanced Driver Assistance Systems (ADAS). This research focuses on a radar-guided YOLOv8s framework to detect vehicles in an FM-CW based 77 GHz radar system and to generate computational traffic situation awareness in India. The upstream radar-processing approach used with it includes range–Doppler processing, Cell-Averaging Constant False Alarm Rate (CA-CFAR) detection, connected-component clustering, and velocity-based motion classification. Timestamps of radar measurements are correlated with car, motorcycle, bus, and truck detection using YOLOv8s algorithms, and processed data are combined. Visual candidates receive a vertical region of interest filtering along with a bounding-box area filtering, and the radar target count is used to apply a rank-sort based on a horizontal proximity to the image centre and on the area of the bounding-box. The experimental evaluation was carried out using 3,099 camera frames and 2,233 radar detection records over 998 nonempty radar frames. The number of targets detected by the radar in each non-empty frame was found to be 2.24 on average and the mean target range was 28.26 m. YOLOv8s was estimated at 3.64 detections per frame with 97.71% detection continuity while radar-guided selection averaged 2.04 detections per frame with 90.30% target fulfillment reported. The radar statistics were checked against processed records, but the camera-performance statistics needed to be reproduced. The framework shows frame-by-frame detection count guided visual selection without establishing the object level radar–camera correspondence. Its outputs are complementary vehicle classification, range and motion information pertinent to traffic perception and downstream decision-support applications for ADAS.
Zitieren
APA
Kumar, U., & S, D. K. (2026). Millimeter Wave Radar Guided Yolov8-Based Camera Perception for Vehicle Detection in Indian Traffic. Computer and Decision Making An International Journal. https://doi.org/10.1922/comdem.v3i6s.18781
Vancouver
Kumar U, S DK. Millimeter Wave Radar Guided Yolov8-Based Camera Perception for Vehicle Detection in Indian Traffic. Computer and Decision Making An International Journal. 2026. doi:10.1922/comdem.v3i6s.18781
BibTeX
@article{Kumar2026,
author = {Kumar, Udhaya and S, Dr. Kusumadhara},
title = {Millimeter Wave Radar Guided Yolov8-Based Camera Perception for Vehicle Detection in Indian Traffic},
journal = {Computer and Decision Making An International Journal},
year = {2026},
doi = {10.1922/comdem.v3i6s.18781}
}
EUVIMED