Fractional B-Spline Wavelets and U-Net Architecture for Robust and Reliable Vehicle Detection in Snowy Conditions
Abstract
:1. Introduction
2. Proposed Method
2.1. Fractional B-Spline Wavelet Transform
2.2. Integration with U-Net
3. Dataset
4. Testing and Evaluation
- YOLOv5s.
- YOLOv8s.
- Faster R-CNN.
4.1. Evaluation Metrics and Benchmarking
- When a pair of bounding boxes (one from the predictions and one from the ground truth) achieves an IoU exceeding a predefined threshold, we classify the prediction as accurate, or a True Positive.
- Should a predicted bounding box fail to meet the IoU threshold with any ground-truth bounding boxes, we categorize the prediction as a False Positive.
- Conversely, if a ground-truth bounding box does not reach the IoU threshold with any predicted bounding boxes, we label the prediction as a False Negative.
4.2. Experimental Results
5. Ethical Considerations
- Accuracy: Emphasis was placed on the high-quality annotations of the dataset, which is specific to snowy conditions in the Nordic region, acknowledging potential limitations in generalizability and the importance of accurate data for training algorithms.
- Transparency: The methodology for data collection and model training is thoroughly documented, promoting scrutiny and validation by the scientific community. The use of deep learning for vehicle detection is well-explained, with intentions to share findings and ensure the algorithms perform as expected without unintended behaviors. The dataset used is publicly available.
- Privacy: The aerial data collection minimizes privacy risks by focusing on vehicle tops, excluding identifiable details like license plates or human faces, ensuring anonymity and compliance with privacy standards.
- Fairness and Bias: The dataset encompasses a diverse range of vehicles under snowy conditions to mitigate bias. Careful dataset splitting ensures balanced training and testing sets, allowing for an equitable assessment of the model’s performance in snow-covered environments.
6. Conclusions and Future Work
Funding
Data Availability Statement
Conflicts of Interest
References
- Xu, Y.; Liu, X.; Cao, X.; Huang, C.; Liu, E.; Qian, S.; Liu, X.; Wu, Y.; Dong, F.; Qiu, C.W.; et al. Artificial intelligence: A powerful paradigm for scientific research. Innovation 2021, 2, 100179. [Google Scholar] [CrossRef] [PubMed]
- Tahir, N.U.A.; Zhang, Z.; Asim, M.; Chen, J.; ELAffendi, M. Object Detection in Autonomous Vehicles under Adverse Weather: A Review of Traditional and Deep Learning Approaches. Algorithms 2024, 17, 103. [Google Scholar] [CrossRef]
- Chen, C.; Wang, C.; Liu, B.; He, C.; Cong, L.; Wan, S. Edge intelligence empowered vehicle detection and image segmentation for autonomous vehicles. IEEE Trans. Intell. Transp. Syst. 2023, 24, 13023–13034. [Google Scholar] [CrossRef]
- Mokayed, H.; Nayebiastaneh, A.; De, K.; Sozos, S.; Hagner, O.; Backe, B. Nordic Vehicle Dataset (NVD): Performance of vehicle detectors using newly captured NVD from UAV in different snowy weather conditions. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Vancouver, BC, Canada, 17–24 June 2023; pp. 5313–5321. [Google Scholar]
- Lee, S.H.; Lee, S.H. U-Net-Based Learning Using Enhanced Lane Detection with Directional Lane Attention Maps for Various Driving Environments. Mathematics 2024, 12, 1206. [Google Scholar] [CrossRef]
- Chen, B.; Fan, X. MSGC-YOLO: An Improved Lightweight Traffic Sign Detection Model under Snow Conditions. Mathematics 2024, 12, 1539. [Google Scholar] [CrossRef]
- Li, S.; Yang, X.; Lin, X.; Zhang, Y.; Wu, J. Real-time vehicle detection from UAV aerial images based on improved YOLOv5. Sensors 2023, 23, 5634. [Google Scholar] [CrossRef] [PubMed]
- Bulut, A.; Ozdemir, F.; Bostanci, Y.S.; Soyturk, M. Performance Evaluation of Recent Object Detection Models for Traffic Safety Applications on Edge. In Proceedings of the 2023 5th International Conference on Image Processing and Machine Vision, Macau, China, 13–15 January 2023; pp. 1–6. [Google Scholar]
- Liu, Z.; Gao, Y.; Du, Q.; Chen, M.; Lv, W. YOLO-extract: Improved YOLOv5 for aircraft object detection in remote sensing images. IEEE Access 2023, 11, 1742–1751. [Google Scholar] [CrossRef]
- Huang, F.; Chen, S.; Wang, Q.; Chen, Y.; Zhang, D. Using deep learning in an embedded system for real-time target detection based on images from an unmanned aerial vehicle: Vehicle detection as a case study. Int. J. Digit. Earth 2023, 16, 910–936. [Google Scholar] [CrossRef]
- Mokayed, H.; Nayebiastaneh, A.; Alkhaled, L.; Sozos, S.; Hagner, O.; Backe, B. Challenging YOLO and Faster RCNN in Snowy Conditions: UAV Nordic Vehicle Dataset (NVD) as an Example. In Proceedings of the 2024 2nd International Conference on Unmanned Vehicle Systems-Oman (UVS), Muscat, Oman, 12–14 February 2024; pp. 1–6. [Google Scholar]
- John, S.M.; Kareem, F.A.; Paul, S.G.; Gafur, A.; Al Mansoori, S.; Panthakkan, A. Enhanced YOLOv7 Model for Accurate Vehicle Detection from UAV Imagery. In Proceedings of the 2023 International Conference on Innovations in Engineering and Technology (ICIET), Muvattupuzha, India, 13–14 July 2023; pp. 1–4. [Google Scholar]
- Javid, I.; Ghazali, R.; Saeed, W.; Batool, T.; Al-Wajih, E. CNN with New Spatial Pyramid Pooling and Advanced Filter-Based Techniques: Revolutionizing Traffic Monitoring via Aerial Images. Sustainability 2023, 16, 117. [Google Scholar] [CrossRef]
- Tanasa, I.Y.; Budiarti, D.H.; Guno, Y.; Yunata, A.S.; Wibowo, M.; Hidayat, A.; Purnamastuti, F.N.; Purwanto, A.; Wicaksono, G.; Domiri, D.D. U-Net Utilization on segmentation of Aerial Captured Images. In Proceedings of the 2023 International Conference on Radar, Antenna, Microwave, Electronics, and Telecommunications (ICRAMET), Bandung, Indonesia, 15–16 November 2023; pp. 107–112. [Google Scholar]
- Mokayed, H.; Shivakumara, P.; Woon, H.H.; Kankanhalli, M.; Lu, T.; Pal, U. A new DCT-PCM method for license plate number detection in drone images. Pattern Recognit. Lett. 2021, 148, 45–53. [Google Scholar] [CrossRef]
- Chen, Z.; Yang, J.; Feng, Z.; Zhu, H. RailFOD23: A dataset for foreign object detection on railroad transmission lines. Sci. Data 2024, 11, 72. [Google Scholar] [CrossRef]
- McKean, H.; Moll, V. Elliptic Curves: Function Theory, Geometry, Arithmetic; Cambridge University Press: Cambridge, UK, 1999. [Google Scholar]
Paper | Model | Edge Device | Dataset/Weather Condition |
---|---|---|---|
Li et al., 2023 [7] | YOLOv5s, YOLOv5 | CPU Intel(R) Xeon(R) Platinum 8358P CPU @ 2.60 GHz, GPU: RTX A5000-24 GB. | VisDrone2019-DET No specific weather conditions. |
Bulut et al., 2023 [8] | YOLOv5, YOLOv7, YOLOv6 | NVIDIA Jetson Nano | No specific weather conditions. |
Liu et al., 2023 [9] | improved version of the YOLOv5 called YOLO-Extract. | X | DOTA dataset. No specific weather conditions. |
Huang et al., 2023 [10] | YOLOv4 | NVIDIA Jetson Nano | No specific weather conditions. |
Mokayed et al., 2024 [11] | YOLOv5s, YOLOv8s, SSD, FRCNN | X | NVD dataset with severe snowy conditions |
John et al., 2023 [12] | YOLOv5 and YOLOv7, with YOLOv7 being the primary model | X | VEDAI dataset |
Javid et al., 2024 [13] | (CNN)-based U-Net model | X | DLR3K and Vedai datasets. |
Tanasa et al., 2023 [14] | U-Net | NVIDIA Quadro RTX 4000 GPUs with 8 GB of memory, and NVIDIA Pascal GPU-based TX2 edge devices with 8 GB of memory. | No specific weather conditions. |
Mokayed et al., 2021 [15] | DCT-PCM with conventional CNN | NUC intel i7 processor without GPU. | Mimos dataset, no specific weather conditions. |
Video | Altitude | Snow Cover | Cloud Cover | fps | GSD |
---|---|---|---|---|---|
Asjo 01 | 130–200 m | minimal (0–1 cm) | overcast | 5 | 11.5–17.8 cm |
Bjenberg | 250 m | Fresh (1–2 cm) | light | 25 | 22.2 cm |
Asjo 01 HD | 250 m | Fresh (5–10 cm) | clear | 5 | 20.2 cm |
Video | Altitude | Snow Cover | Cloud Cover | fps | GSD |
---|---|---|---|---|---|
Bjenberg 02 | 250 m | Fresh (5–10 cm) | clear | 5 | 11.1cm |
Nyland-01 | 150 m | Minimal (0–1 cm) | Dense | 5 | 11.5–17.8 cm |
Part | Specification |
---|---|
Processor | Intel i9-9900K @ 3.6 GHz |
RAM | 64 GB (3600 MHz) DDR4 CL16 |
Graphic Card | Nvidia Geforce RTX 3800Ti 12 GB (Cuda 11.1) |
Model | Precision | Recall | mAP50 | mAP50-95 |
---|---|---|---|---|
YOLOv5s | 54.2% | 33.7% | 47.3% | 30.5% |
YOLOv8s | 65.8% | 22.4% | 45.1% | 29.8% |
YOLOv5s_aug* | 70.6% | 48.2% | 56.0% | 24.1% |
YOLOv58s_aug* | 77.1% | 34.6% | 50.7% | 24.1% |
Faster RCNN (FPN3x) | - | - | 46.2% | - |
CarLocalizationCNN | 74.7% | 54.8% | 60.4% | 38.5% |
Model | Pre-Process (ms) | Inference (ms) | Post-Process (ms) | Total (ms) |
---|---|---|---|---|
YOLOv5s | 52.6 | 17,731.7 | 2.6 | 17,786.9 |
YOLOv8s | 55.5 | 12,137.4 | 3.7 | 12,196.6 |
Faster RCNN (FPN3x) | - | 62,715.87 | - | 62,715.87 |
CarLocalizationCNN | 0.032 | 9348.062 | 361.172 | 9709.266 |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Mokayed, H.; Ulehla, C.; Shurdhaj, E.; Nayebiastaneh, A.; Alkhaled, L.; Hagner, O.; Hum, Y.C. Fractional B-Spline Wavelets and U-Net Architecture for Robust and Reliable Vehicle Detection in Snowy Conditions. Sensors 2024, 24, 3938. https://doi.org/10.3390/s24123938
Mokayed H, Ulehla C, Shurdhaj E, Nayebiastaneh A, Alkhaled L, Hagner O, Hum YC. Fractional B-Spline Wavelets and U-Net Architecture for Robust and Reliable Vehicle Detection in Snowy Conditions. Sensors. 2024; 24(12):3938. https://doi.org/10.3390/s24123938
Chicago/Turabian StyleMokayed, Hamam, Christián Ulehla, Elda Shurdhaj, Amirhossein Nayebiastaneh, Lama Alkhaled, Olle Hagner, and Yan Chai Hum. 2024. "Fractional B-Spline Wavelets and U-Net Architecture for Robust and Reliable Vehicle Detection in Snowy Conditions" Sensors 24, no. 12: 3938. https://doi.org/10.3390/s24123938
APA StyleMokayed, H., Ulehla, C., Shurdhaj, E., Nayebiastaneh, A., Alkhaled, L., Hagner, O., & Hum, Y. C. (2024). Fractional B-Spline Wavelets and U-Net Architecture for Robust and Reliable Vehicle Detection in Snowy Conditions. Sensors, 24(12), 3938. https://doi.org/10.3390/s24123938