Application of Image Processing Techniques for UAV Detection Using Deep Learning and Distance-Wise Analysis in an open-access work by Bedanshu Dewangan, Aditya Saxena, Rahul Thakur, and Shrivishal Tripath.
Drones are increasingly integrated into various aspects of daily life, serving purposes in agriculture, military operations, commercial ventures, disaster relief, research, and more. The proliferation of small drones or unmanned aerial vehicles (UAVs) has surged in recent years, raising concerns about their potential misuse for illegal activities, including terrorism and drug trafficking. This underscores the urgent need for accurate and reliable UAV identification systems capable of functioning across diverse environments.
This paper explores different iterations of the state-of-the-art YOLO (You Only Look Once) object detection models, which leverage computer vision and deep learning techniques to detect small UAVs. The authors propose integrating various image-processing methods into the existing YOLO models to enhance detection accuracy, resulting in notable performance improvements.
This study achieved a mean Average Precision (mAP) score of 96.7%, an Intersection over Union (IoU) threshold of 50%, a precision of 95%, and a recall of 95.6%. Additionally, we conducted a distance-wise analysis, evaluating detection accuracy for UAVs at close, mid, and far ranges.
Publication Date– March 2023
Application of Image Processing Techniques for UAV Detection Using Deep Learning and Distance-Wise Analysis contains the following major sections:
- Introduction
- Related Work and Contributions
- Methodology
- Experimental Setup
- Results and Explanation
- Conclusions
Post Image Credit- Adobe Stock by Di Studio (Made with AI)
