In the field of intelligent transportation systems and urban security, it is crucial to obtain accurate information of vehicles. Vehicle-related information can be directly obtained through visual recognition means such as video or images. However, in low-light environments, the image brightness and contrast decrease, the noise level increases, and the image features are prone to loss. These problems lead to a significant reduction in the detection accuracy of vehicle detection algorithms. Therefore, we propose a vehicle detection method based on low-light image enhancement and an improved object detection algorithm. The low-light image was first enhanced using the image enhancement algorithm ZeroDCE to improve the image brightness. Then, the improved AFF-YOLO object detection algorithm is utilized to perform vehicle detection on the enhanced image. Finally, the proposed method is tested on a vehicle dataset, and the vehicle detection accuracy under different low-light levels is analyzed. The results show that the proposed method can effectively improve the vehicle detection accuracy. Compared with low-light images, mAP@0.5 of the enhanced images improved by 4.9% to 94.7%. As the illumination intensity decreases, the object detection accuracy of the enhanced image improves more significantly. The research results can provide a reference for vehicle detection in low-light environments.
为了说明本文方法的优势,首先对比了8种不同增强算法的性能,从增强算法对目标检测精度的提升效果、算法的参数量和运行速度三个方面进行对比,结果如表3所示. 其中,目标检测精度指标选用交并比(Intersection over Union,IoU)为0.5时的平均精度均值(Mean Average Precision,mAP)[26]. 由表3可知,不论是采用YOLOv7或AFF-YOLO进行目标检测,ZeroDCE的增强效果在所有方法中均相对较好,并且ZeroDCE具有参数量少、计算速度快等优势.
WANGY F, ZHUX R, YUNX,et al .Vehicle re-identification by multi-cameras for public security surveillance[J].Journal of Xidian University,2019,46(4):190-196.(in Chinese)
QIANL .Design and implementation of an urban high-emission vehicle restriction management system for intelligent transportation[D].Wuhan:Huazhong University of Science and Technology,2022:1-6.(in Chinese)
[5]
ZAURINR, KHUCT, CATBASF N .Hybrid sensor-camera monitoring for damage detection:case study of a real bridge[J].Journal of Bridge Engineering,2016,21(6):05016002.
[6]
SHUM, ZHONGY F, LÜP Y .Small moving vehicle detection via local enhancement fusion for satellite video[J].International Journal of Remote Sensing,2021,42(19):7189-7214.
[7]
CHENZ C, LIH, BAOY Q,et al .Identification of spatio-temporal distribution of vehicle loads on long-span bridges using computer vision technology[J].Structural Control and Health Monitoring,2016,23(3):517-534.
[8]
SUNZ, BEBISG, MILLERR .On-road vehicle detection using optical sensors:a review[C]//Proceedings of The 7th International IEEE Conference on Intelligent Transportation Systems.Washington,WA,USA: IEEE,2004:585-590.
[9]
CHENZ W, FENGY L, ZHANGY,et al .An accurate and convenient method of vehicle spatiotemporal distribution recognition based on computer vision[J].Sensors,2022,22(17):1-17.
[10]
ZHANGB, ZHOUL M, ZHANGJ .A methodology for obtaining spatiotemporal information of the vehicles on bridges based on computer vision[J].Computer-Aided Civil and Infrastructure Engineering,2019,34(6):471-487.
[11]
GEL F, DAND H, KOOK Y,et al .An improved system for long-term monitoring of full-bridge traffic load distribution on long-span bridges[J].Structures,2023,54:1076-1089.
[12]
ZHOUY, PEIY L, LIZ W,et al .Vehicle weight identification system for spatiotemporal load distribution on bridges based on non-contact machine vision technology and deep learning algorithms[J].Measurement,2020,159:107801.
ZHOUY, HUJ N, ZHAOY,et al .Vehicle target tracking based on Kalman filtering improved compressed sensing algorithm[J].Journal of Hunan University (Natural Sciences), 2023, 50(1):11-21.(in Chinese)
[15]
ZHUJ S, LIX T, ZHANGC,et al .An accurate approach for obtaining spatiotemporal information of vehicle loads on bridges based on 3D bounding box reconstruction with computer vision[J]. Measurement, 2021, 181: 109657.
[16]
胡忆平 .基于深度学习的夜间车辆检测算法研究[D].长沙:湖南大学,2022:3-8.
[17]
HUY P .Research on night vehicle detection algorithm based on deep learning[D].Changsha:Hunan University,2022:3-8.(in Chinese)
[18]
WANGY, XIEW J, LIUH Q .Low-light image enhancement based on deep learning:a survey[J].Optical Engineering,2022,61:040901.
[19]
SOBBAHI RAL, TEKLIJ .Comparing deep learning models for low-light natural scene image enhancement and their impact on object detection and classification:overview,empirical evaluation,and challenges[J].Signal Processing:Image Communication,2022, 109:116848.
MAL, MAT Y, LIUR S .The review of low-light image enhancement[J].Journal of Image and Graphics,2022,27(5):1392-1409.(in Chinese)
[22]
ABDULLAH-AL-WADUDM, KABIRM H, AKBER DEWAN MALI,et al .A dynamic histogram equalization for image contrast enhancement[J]. IEEE Transactions on Consumer Electronics, 2007, 53(2): 593-600.
[23]
HUANGS C, CHENGF C, CHIUY S .Efficient contrast enhancement using adaptive gamma correction with weighting distribution[J].IEEE Transactions on Image Processing, 2013,22(3): 1032-1041.
[24]
LANDE H .The retinex theory of color vision[J]. Scientific American, 1977, 237(6): 108-128.
[25]
LIC Y, GUOC L, HANL H,et al .Low-light image and video enhancement using deep learning:a survey[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2022,44(12):9396-9416.
[26]
YUR S, LIUW Y, ZHANGY S,et al .DeepExposure:learning to expose photos with asynchronously reinforced adversarial learning[C]//Proceedings of the 32nd International Conference on Neural Information Processing Systems.Montréal,Canada:ACM,2018:2153-2163.
[27]
GUOC L, LIC Y, GUOJ C,et al .Zero-reference deep curve estimation for low-light image enhancement[C]// Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Seattle,WA,USA:IEEE, 2020: 1777-1786.
[28]
LIC Y, GUOC L, LOYC C .Learning to enhance low-light image via zero-reference deep curve estimation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2022,44(8):4225-4238.
[29]
WANGC Y, BOCHKOVSKIYA, LIAOH Y M .YOLOv7:trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[C]//Proceedings of the 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Vancouver,BC,Canada:IEEE,2023:7464-7475.
[30]
WENL Y, DUD W, CAIZ W,et al .UA-DETRAC:a new benchmark and protocol for multi-object detection and tracking[J].Computer Vision and Image Understanding,2020,193:102907.
[31]
PADILLAR, NETTOS L, SILVA E A BDA .A survey on performance metrics for object-detection algorithms[C]//2020 International Conference on Systems,Signals and Image Processing (IWSSIP).Niteroi,Brazil:IEEE,2020:237-242.
基金资助
国家自然科学基金资助项目(52008160)
National Natural ScienceFoundation of China(52008160)
湖南省优秀青年基金项目(2021JJ20015)
Scientific Foundation for Excellent Young Scholars of Hunan Province(2021JJ20015)