Adaptive Low-Light Enhancement for Road Anomaly Detection: A Multi-Architecture Benchmark and Edge Deployment Study

Obed Honour Eje; Oluwatosin Kolade; Glory Bagai; Taliha Abiodun Folorunso

cs.CVNov 5, 2026

Adaptive Low-Light Enhancement for Road Anomaly Detection: A Multi-Architecture Benchmark and Edge Deployment Study

Authors: Obed Honour Eje, Oluwatosin Kolade, Glory Bagai, Taliha Abiodun Folorunso

Organizations: Mechatronics Engineering, Federal University of Technology Minna, Minna, Nigeria · Mechanical Engineering, Obafemi Awolowo University, Ile-Ife, Nigeria

Abstract

Road anomaly detection systems on autonomous vehicles must function reliably under degraded visual conditions including darkness, rain, and Harmattan dust haze. This paper benchmarks four object detection architectures, YOLOv8n, YOLOv12n, RT-DETR, and Faster R-CNN, under physically grounded lighting and weather degradation, and shows that uniform application of the SCI low-light enhancement network is actively harmful at dusk-level lighting (-16.1% to -30.7% mAP@50). We propose a lightweight, training-free luminance-adaptive router that selects between no enhancement, moderate SCI, and denoising plus strong SCI based on per-frame Rec. 601 luminance, recovering most of the dusk penalty while preserving low-light gains, and we validate the full pipeline on NVIDIA Jetson Nano edge hardware.

Figures & tables

Fig. 1: The proposed luminance-adaptive routing architecture.

Model	System	Clean	Dusk	Night	Severe	Dusk D	Night D	Severe D
RT-DETR	Baseline	0.883	0.861	0.612	0.324	–	–	–
	Plain SCI	0.883	0.722	0.679	0.387	-0.139	+0.067	+0.063
	Adaptive	0.883	0.847	0.657	0.338	-0.014	+0.045	+0.014
YOLOv12n	Baseline	0.894	0.867	0.570	0.305	–	–	–
	Plain SCI	0.894	0.706	0.601	0.368	-0.161	+0.031	+0.063
	Adaptive	0.894	0.829	0.577	0.294	-0.038	+0.007	-0.011

TABLE I: Complete mAP@50 Results. D Columns Show Delta vs Baseline

Fig. 2: Baseline mAP@50 under lighting degradation.

Fig. 3: Architecture robustness: mAP@50 retention (percentage of clean baseline) across all lighting and weather conditions.

Model	Branch	Lum.	Enh.	Det.	FPS
		(ms)	(ms)	(ms)
RT-DETR	A (Clean)	1.2	0	120.4	8.23
	B (Dusk)	1.2	8.3	121.3	7.65
	C (Severe)	1.2	27.3	120.9	6.69
YOLOv12n	A (Clean)	1.2	0	52.7	18.54
	B (Dusk)	1.2	13.2	52.4	14.97

TABLE II: Jetson Nano Pipeline Latency. FPS Thresholds:

\geq

20 real-time,

\geq

15 safety minimum,

<

15 below safety

Object Detection

Object detection locates and classifies objects in images or video, balancing localization accuracy, robustness, and inference speed.

See all 648 papers

CADENet: Condition-Adaptive Asynchronous Dual-Stream Enhancement Network for Adverse Weather Perception in Autonomous Driving
May 19, 2026Sherif Khairy, Catherine M. EliasImage Enhancement Autonomous Driving Perception
Computer Science & Engineering Department, German University in Cairo (GUC), Egypt · C-DRiVeS Lab: Cognitive Driving Research in Vehicular Systems, Cairo, Egypt
IAF-Net: Illumination-Adaptive Fusion for Low-Light Urban Road Segmentation
May 29, 2026Bingtao Wang, Daojie Peng, Fulong Ma +2Autonomous Driving Perception Multimodal Fusion
The Shandong University · The Hong Kong University of Science and Technology (Guangzhou)
Self-DACE++: Robust Low-Light Enhancement via Efficient Adaptive Curve Estimation
Apr 28, 2026Jianyu Wen, Jun Xie, Feng Chen +5Image Enhancement Low-Light Image Enhancement
Lenovo Research · Imperial College London · Southern University of Science and Technology +2

May 19, 2026J/K move · Enter open · S save

CADENet: Condition-Adaptive Asynchronous Dual-Stream Enhancement Network for Adverse Weather Perception in Autonomous Driving

Sherif Khairy, Catherine M. Elias

Computer Science & Engineering Department, German University in Cairo (GUC), Egypt · C-DRiVeS Lab: Cognitive Driving Research in Vehicular Systems, Cairo, Egypt

Image Enhancement Autonomous Driving Perception Object Detection Real-Time Object Detection

Open paper AskPDF arXiv

Adverse weather (rain, fog, sand, and snow) degrades camera-based object detection in autonomous vehicles. Existing enhancement-then-detect approaches stall the safety-critical perception loop, violating hard real-time requirements. Progress on this problem is also constrained by an under-recognized evaluation ceiling: ground truth annotated on degraded images cannot credit a detector that recovers objects the annotators themselves could not see, so a genuinely useful enhancement can register as a near-flat F1 gain. This paper presents CADENet (Condition-Adaptive Asynchronous Dual-stream Enhancement Network), a training-free three-thread system: Thread S (YOLOv11n) delivers detections at full frame rate with zero added latency; Thread Q applies condition-adaptive enhancement (CAPE) and fuses results via entropy-guided NMS (EG-NMS) without blocking Thread S; Thread E provides CLIP zero-shot weather classification, so new weather categories require only a new text prompt, with no labeled data and no retraining. Evaluated on 1327 DAWN images (YOLOv11m, IoU = 0.5, confidence = 0.25), CADENet achieves Recall = 0.0103 (micro), F1 = 0.0230 on snow, and F1 = 0.0038 on rain. We formalize the annotation completeness bias on DAWN-class data, so the reported F1 values are lower bounds on the true gain; recall is the annotation-gap-immune headline metric. Thread S sustains approximately 44 FPS regardless of enhancement load. No model retraining or additional sensor hardware is required.

Abstract

Object Detection

Explore similar work

CADENet: Condition-Adaptive Asynchronous Dual-Stream Enhancement Network for Adverse Weather Perception in Autonomous Driving

IAF-Net: Illumination-Adaptive Fusion for Low-Light Urban Road Segmentation

Self-DACE++: Robust Low-Light Enhancement via Efficient Adaptive Curve Estimation