A Comparison of SR and CBAM for Optimized Thermal Drone Object Detection

Helfy Susilawati; Akhmad Fauzi Ikhsan; Firman; Arief Suryadi Satyawan; Chandra Rahmana

doi:10.22146/jnteti.v15i2.24931

Helfy Susilawati Department of Electrical Engineering, Faculty of Engineering, Universitas Garut, Garut, Jawa Barat 44116, Indonesia
Akhmad Fauzi Ikhsan Department of Electrical Engineering, Faculty of Engineering, Universitas Garut, Garut, Jawa Barat 44116, Indonesia
Firman Department of Electrical Engineering, Faculty of Engineering, Universitas Garut, Garut, Jawa Barat 44116, Indonesia
Arief Suryadi Satyawan Research Center for Telecommunication, National Research and Innovation Agency, Bandung, Jawa Barat 40135, Indonesia
Chandra Rahmana Department of Electrical Engineering, Faculty of Engineering, Universitas Garut, Garut, Jawa Barat 44116, Indonesia

DOI: https://doi.org/10.22146/jnteti.v15i2.24931

Keywords: SAR, YOLOv8, Super Resolution, CBAM, Object Detection

Abstract

Human detection using thermal cameras is very useful in certain conditions, such as detecting people lost in mountainous areas that are difficult to explore. Rescue operations are usually conducted by deploying a search and rescue (SAR) team to the location, which is not always effective because this operation can only be carried out under certain conditions and may pose a risk to the SAR team itself. Therefore, one alternative approach is the use of drones equipped with human detection and recognition capabilities. In this context, thermal cameras are used because they can penetrate challenging environments, making them suitable for SAR operations. The object detection method used in this study was You Only Look Once (YOLO) version 8 or YOLOv8. This study aimed to compare the effectiveness of integrating enhanced super-resolution generative adversarial networks (ESRGAN) with YOLOv8 and incorporating a convolutional block attention module (CBAM) into the neck architecture of YOLOv8. The performance of ESRGAN with YOLOv8 and CBAM with YOLOv8 was evaluated using precision, mean average precision (mAP), and training loss. Based on the experimental results, the combination of ESRGAN with YOLOv8 outperformed the CBAM-based modification. This is indicated by higher precision and mAP values, as well as lower training loss in the ESRGAN-enhanced YOLOv8 detection framework. The experimental findings highlight that image enhancement using ESRGAN is more effective than CBAM-based modification in improving thermal image-based human detection performance for SAR applications.

References

Krisdianto, “Pencarian pendaki yang hilang di Gunung Binaya dilanjutkan.” Kementerian Kehutanan RI. Access date: 29-Sep-2025. [Online]. Available: https://www.kehutanan.go.id/news/pencarian-pendaki-yang-hilang-di-gunung-binaya-dilanjutkan

M.B. Ibrahim, “Pemuda Sidoarjo hilang saat mendaki Gunung Buthak.” Detikjatim. Access date: 29-Sep-2025. [Online]. Available: https://www.detik.com/jatim/berita/d-8068473/pemuda-sidoarjo-hilang-saat-mendaki-gunung-buthak

M. Erdelj, E. Natalizio, K.R. Chowdhury, and I.F. Akyildiz, “Help from the sky: Leveraging UAVs for disaster management,” IEEE Pervasive Comput., vol. 16, no. 1, pp. 24–32, Jan.-Mar. 2017, doi: 10.1109/MPRV.2017.11.

Meenakshi et al., “AI-driven autonomous robots for search and rescue operations in disaster zones,” in 2025 Int. Conf. Data Sci. Agents Artif. Intell. (ICDSAAI), 2025, pp. 1–6, doi: 10.1109/ICDSAAI65575.2025.11011570.

F. Mercogliano et al., “ITINERIS geophysical technologies @CNR-IREA: Drone-based tests at Altopiano di Verteglia, Avellino (Southern Italy),” Comput. Sci. Appl. – ICCSA 2025 Workshops, 2025, pp. 373–383, doi: 10.1007/978-3-031-97663-6_33.

H. Shi et al., “Improving surface soil moisture content estimation of winter oilseed rape by integrating multimodal remote sensing information and structural characteristics,” Comput. Electron. Agric., vol. 237, pp. 1–18, Oct. 2025, doi: 10.1016/j.compag.2025.110779.

T. Sam-Odusina et al., “Detection and geolocation of peat fires using thermal infrared cameras on drones,” Drones, vol. 9, no. 7, pp. 1–19, Jul. 2025, doi: 10.3390/drones9070459.

B. Chang et al., “Enhancing wildlife detection using thermal imaging drones: Designing the flight path,” Drones, vol. 9, no. 1, pp. 1–14, Jan. 2025, doi: 10.3390/drones9010052.

S. Yeom, “Thermal image tracking for search and rescue missions with a drone,” Drones, vol. 8, no. 2, pp. 1–15, Feb. 2024, doi: 10.3390/drones8020053.

S. Yeom, “Multi-rotor drone-based thermal target tracking with track segment association for search and rescue missions,” Drones, vol. 8, no. 11, pp. 1–18, Nov. 2024, doi: 10.3390/drones8110689.

H. Farman et al., “Enhancing UAV-based human detection in thermal imaging with custom dataset,” in Proc. ICSDI 2024 Vol. 3, 2024, pp. 437–444, doi: 10.1007/978-981-97-8345-8_53.

J. McGee, S.J. Mathew, and F. Gonzalez, “Unmanned aerial vehicle and artificial intelligence for thermal target detection in search and rescue applications,” in 2020 Int. Conf. Unmanned Aircr. Syst. (ICUAS), 2020, pp. 883–891, doi: 10.1109/ICUAS48674.2020.9213849.

V. Teju and D. Bhavana, “An efficient object tracking in thermal imaging using optimal Kalman filter,” Int. J. Eng. Trends Technol., vol. 69, no. 12, pp. 197–202, Dec. 2021, doi: 10.14445/22315381/IJETT-V69I12P223.

L. Courtrai, M.T. Pham, and S. Lefèvre, “Small object detection in remote sensing images based on super-resolution with auxiliary generative adversarial networks,” Remote Sens., vol. 12, no. 19, pp. 1–19, Sep. 2020, doi: 10.3390/rs12193152.

I. García-Aguilar, J. García-González, R.M. Luque-Baena, and E. López-Rubio, “Object detection in traffic videos: An optimized approach using super-resolution and maximal clique algorithm,” Neural Comput. Appl., vol. 35, no. 26, pp. 18999–19013, Sep. 2023, doi: 10.1007/s00521-023-08741-4.

Y.R. Musunuri, O.-S. Kwon, and S.-Y. Kung, “SRODNet: Object detection network based on super resolution for autonomous vehicles,” Remote Sens., vol. 14, no. 24, pp. 1–19, Dec. 2022, doi: 10.3390/rs14246270.

S. Ren et al., “Towards efficient video detection object super-resolution with deep fusion network for public safety,” Secur. Commun. Networks, vol. 2021, pp. 1–14, May 2021, doi: 10.1155/2021/9999398.

Y. Zhang et al., “Research of maritime object detection method in foggy environment based on improved model SRC-YOLO,” Sensors, vol. 22, no. 20, pp. 1–14, Oct. 2022, doi: 10.3390/s22207786.

W. Wang et al., “Improved detection and reidentification algorithm for natural gas pipeline excavator dynamic tracking based on deep learning,” Measurement, vol. 257, pp. 1–17, Jan. 2026, doi: 10.1016/j.measurement.2025.118567.

Z. Lv, S. Zhu, D. Wang, and Z. Liang, “Infrared-visible person re-identification via dual-channel attention mechanism,” Multimed. Tools Appl., vol. 82, no. 15, pp. 22631–22649, Jun. 2023, doi: 10.1007/s11042-023-14486-y.

C. Song et al., “Smoke video detection based on double spectrum,” Int. J. Robot. Autom., vol. 39, no. 2, pp. 161–169, Jan. 2024, doi: 10.2316/J.2024.206-1020.

M. Rizk and I. Bayad, “Bringing intelligence to SAR missions: A comprehensive dataset and evaluation of YOLO for human detection in TIR images,” IEEE Access, vol. 13, pp. 17208–17235, Jan. 2025, doi: 10.1109/ACCESS.2025.3529484.

J. Dong, K. Ota, and M. Dong, “UAV-based real-time survivor detection system in post-disaster search and rescue operations,” IEEE J. Miniaturization Air Space Syst., vol. 2, no. 4, pp. 209–219, Dec. 2021, doi: 10.1109/JMASS.2021.3083659.

Z. Song et al., “An infrared dataset for partially occluded person detection in complex environment for search and rescue,” Sci. Data, vol. 12, pp. 1–14, Feb. 2025, doi: 10.1038/s41597-025-04600-0.

G. Batchuluun et al., “Deep learning-based thermal image reconstruction and object detection,” IEEE Access, vol. 9, pp. 5951–5971, Dec. 2020, doi: 10.1109/ACCESS.2020.3048437.

F. Ponzini, D. van Hamme, and M. Martelli, “Human detection in marine disaster search and rescue scenario: A multi-modal early fusion approach,” Ocean Eng., vol. 340, pp. 1–15, Nov. 2025, doi: 10.1016/j.oceaneng.2025.122341.

A. Vaswani et al., “Attention is all you need,” 2023, arXiv:1706.03762.

P. Shetty, D.K.R, and D.S. Padre, “Deepfake detection using hybrid deep learning: Enhancing accuracy with ResNet, GANs, and attention mechanisms,” in 2025 3rd Int. Conf. Augment. Intell. Sustain. Syst. (ICAISS), 2025, pp. 95–99, doi: 10.1109/ICAISS61471.2025.11041976.

F. Ying et al., “Improvements in automatic diagnosis methods for knee osteoarthritis based on deep learning,” Chinese J. Tissue Eng. Res., vol. 29, no. 35, pp. 7511–7518, 2025, doi: 10.12307/2026.533.

J. Wang, J. Gao, and B. Zhang, “A small object detection model in aerial images based on CPDD-YOLOv8,” Sci. Rep., vol. 15, pp. 1–16, Jan. 2025, doi: 10.1038/s41598-024-84938-4.

Y. Gao and B. Lin, “Research and design of a small object detection and tracking system based on YOLOv8 and ByteTrack algorithms,” in 2024 IEEE 4th Int. Conf. Data Sci. Comput. Appl. (ICDSCA), 2024, pp. 109–113, doi: 10.1109/ICDSCA63855.2024.10859594.

H. Zhang, M. Liang, and Y. Wang, “YOLO-BS: A traffic sign detection algorithm based on YOLOv8,” Sci. Rep., vol. 15, pp. 1–11, Mar. 2025, doi: 10.1038/s41598-025-88184-0.

R. Ma et al., “InSAR-YOLOv8 for wide-area landslide detection in InSAR measurements,” Sci. Rep., vol. 15, pp. 1–22, Jan. 2025, doi: 10.1038/s41598-024-84626-3.

J. Li et al., “A small-scale target enhancement framework for aerial pineapple images on accurate agricultural information,” Comput. Electron. Agric., vol. 239, pp. 1–14, Dec. 2025, doi: 10.1016/j.compag.2025.110874.

Y. Chen, J. Li, Y. Niu, and J. He, “Small object detection networks based on classification-oriented super-resolution GAN for UAV aerial imagery,” in 2019 Chin. Control Decis. Conf. (CCDC), 2019, pp. 4610–4615, doi: 10.1109/CCDC.2019.8832735.

C. Ledig et al., “Photo-realistic single image super-resolution using a generative adversarial network,” 2016, arXiv:1609.04802.

L. Riz et al., “The MONET dataset: Multimodal drone thermal dataset recorded in rural scenarios,” 2023, arXiv:2304.05417.

X. Wang et al., “ESRGAN: Enhanced super-resolution generative adversarial networks,” 2019, arXiv:1809.00219.

S. Woo, J. Park, J. Lee, and I.S. Kweon, “CBAM: Convolutional block attention module,” 2018, arXiv:1807.06521.

H.G. Correa et al., “Thermal image super-resolution using real-ESRGAN for human detection,” in Proc. 20th Int. Jt. Conf. Comput. Vis. Imaging Comput. Graph. Theory Appl., 2025, pp. 247–254, doi: 10.5220/0013078800003912.