STM Journals Peer-reviewed research from Consortium e-Learning Network

Journal of Polymer & Composites Original Research

Polymer Composite-Enabled UAV Platform for Edge AI-Based Precision Agriculture: A System-Level Evaluation

  1. Patan Imran Khan Department of Electronics and Communication Engineering, St. Johns College of Engineering and Technology, Yemmiganur, Kurnool(D)
  2. Y Narasimha Reddy Department of Computer Science and Engineering, St. Johns College of Engineering and Technology(A), Yemmiganur, Kurnool(D)
  3. P Veeresh Department of Computer Science and Engineering, St. Johns College of Engineering and Technology(A), Yemmiganur, Kurnool(D)
  4. Akhtar Khan Department of Mechanical Engineering, Indian Institute of Information Technology, Design and Manufacturing, Kurnool

Abstract

This study investigates the system-level role of commercially available polymer composite materials in enabling lightweight and energy-efficient unmanned aerial vehicle (UAV) platforms integrated with edge artificial intelligence for real-time agricultural monitoring. Rather than developing or experimentally characterizing new composite materials, the work evaluates fiber-reinforced polymer (FRP) composites and epoxy-based laminates as enabling structural components whose established properties support UAV performance in precision agriculture. Their high strength-to-weight ratio, corrosion resistance, and vibration damping characteristics contribute to improved flight endurance, sensing stability, and energy efficiency. Polymer composites used in UAV frames, FR4 epoxy-glass printed circuit boards, and protective housings provide lightweight construction and mechanical reliability essential for sustained aerial operations. The proposed system, AgroVision-Edge, is deployed on a 1.4 kg quadrotor equipped with a Raspberry Pi 4B (8 GB), Sony IMX477 imaging sensor, and u-blox NEO-M9N GNSS module. Its classification backbone, AgroVision-Net, extends MobileNetV3-Large with a lightweight channel-attention mechanism and is trained on a geo-referenced dataset of 23,200 aerial images covering five crop disease categories. Post-training INT8 quantisation reduces the model size to 1.87 MB while enabling efficient execution on embedded ARM platforms. Experimental results demonstrate a top-1 classification accuracy of 98.1% on a plot-disjoint test set of 4,640 images, with a mean end-to-end inference latency of 43.3 ms and a 95th-percentile latency of 108 ms under concurrent workloads. A geo-referenced Coverage Score of 97.7% over a 4.2-hectare field confirms high spatial detection fidelity. The lightweight polymer-based UAV structure contributes to reduced power consumption of 8.4 W and sustained flight endurance, enabling continuous real-time inference without cloud dependence. The proposed framework demonstrates the effective integration of polymer-based UAV systems and edge AI for scalable, energy-efficient precision agriculture in resource-constrained environments.

Keywords

References (23)

  1. Food and Agriculture Organization of the United Nations, The State of Food and Agriculture 2023. Rome, Italy: FAO, 2023.
  2. D. G. Gilchrist and R. T. Rajaram, “Epidemiological thresholds for early fungal disease intervention in Triticum aestivum,” Plant Pathology Journal, vol. 71, no. 4, pp. 318–329, 2022.
  3. M. A. Torres and P. K. Sharma, “Ground sampling distance optimization for canopy-scale disease detection in UAV multispectral surveys,” Precision Agriculture, vol. 24, no. 2, pp. 612–631, 2023.
  4. C. J. Liu, Y. Zhao, and H. Wang, “Latency analysis of cloud-based deep learning inference for real-time agricultural UAV applications,” IEEE Access, vol. 11, pp. 44218–44229, 2023.
  5. W. Shi, J. Cao, Q. Zhang, Y. Li, and L. Xu, “Edge computing: Vision and challenges,” IEEE Internet of Things Journal, vol. 3, no. 5, pp. 637–646, Oct. 2016.
  6. S. Han, H. Mao, and W. J. Dally, “Deep compression: Compressing deep neural networks with pruning, trained quantization and Huffman coding,” in Proc. Int. Conf. Learning Representations (ICLR), San Juan, PR, USA, 2016.
  7. S. Sankaran, A. Mishra, R. Ehsani, and C. Davis, “A review of advanced techniques for detecting plant diseases,” Computers and Electronics in Agriculture, vol. 72, no. 1, pp. 1–13, 2010.
  8. J. G. A. Barbedo, “A review on the main challenges in automatic plant disease identification using visible spectrum images,” Biosystems Engineering, vol. 144, pp. 52–60, 2016.
  9. X. Zhang, C. Zhao, and Y. Ma, “MobileNetV2-based leaf disease detection on Jetson Nano for edge agricultural applications,” Biosystems Engineering, vol. 218, pp. 115–126, 2022.
  10. Y. Chen, H. Zhou, X. Li, and G. Xue, “Deep convolutional neural networks for wheat rust and blight classification using UAV RGB imagery,” Remote Sensing, vol. 14, no. 8, p. 1923, 2022.
  11. S. Roy, A. Das, and P. Kumar, “Detection of maize northern leaf blight using UAV multispectral imagery and ensemble deep learning,” Computers and Electronics in Agriculture, vol. 205, p. 107639, 2023.
  12. A. Howard et al., “Searching for MobileNetV3,” in Proc. IEEE/CVF Int. Conf. Computer Vision (ICCV), Seoul, South Korea, pp. 1314–1324, 2019.
  13. Z. Huang, R. Lin, and T. Nguyen, “Cloud-assisted ResNet-50 for large-scale wheat rust detection using UAV multispectral imagery,” IEEE Transactions on Geoscience and Remote Sensing, vol. 61, p. 4706815, 2023.
  14. P. Alvarez and M. Santos, “EfficientNet-B3 for airborne maize disease detection using cloud-assisted inference,” Computers and Electronics in Agriculture, vol. 212, p. 108072, 2023.
  15. V. Kumar, N. Singh, and A. Kaur, “Real-time crop disease classification using MobileNetV2 on UAV platforms,” Journal of Field Robotics, vol. 40, no. 3, pp. 521–537, 2023.
  16. S. Park and J. Lee, “SqueezeNet-based onboard crop disease detection with geo-tagging using UAV systems,” Precision Agriculture, vol. 23, no. 4, pp. 1482–1501, 2022.
  17. G. Li, T. Wu, and Q. Zhou, “Edge TPU-accelerated MobileNet for precision orchard disease monitoring,” IEEE Sensors Journal, vol. 23, no. 10, pp. 11208–11219, 2023.
  18. Y. Chen, B. Luo, and W. Zhang, “On-device crop disease classification using MobileNetV2 on Raspberry Pi 4 with 4G communication,” Computers and Electronics in Agriculture, vol. 198, p. 107062, 2022.
  19. M. Weiss and F. Jacob, “Remote sensing applications in agriculture: A comprehensive review,” Remote Sensing of Environment, vol. 236, p. 111402, 2020.
  20. K. L. Patel and A. Gupta, “Georeferenced UAV imagery for precision disease management in dryland wheat systems,” Computers and Electronics in Agriculture, vol. 198, p. 107024, 2022.
  21. V. R. A. Prakash et al., “Synthesis and characterization of lightweight UAV composite materials,” Biomass Conversion and Biorefinery, 2024.
  22. M. M. ElFaham et al., “UAV composite materials characterization,” Journal of Molecular Structure, 2020.
  23. S. Sonkar et al., “Composite UAV airframe development,” Journal of the Brazilian Society of Mechanical Sciences, 2024.
Support