EKF-Based Depth Camera and Deep Learning Fusion for UAV-Person Distance Estimation and Following in SAR Operations
#UAV#Search and Rescue#Depth Camera#Deep Learning#Distance Estimation#Extended Kalman Filter#YOLO-pose#Autonomous Systems
📌 Key Takeaways
Researchers developed a UAV system for maintaining safe distances from people in search and rescue operations
The system fuses depth camera data with monocular camera-to-body distance estimation using YOLO-pose and Extended Kalman Filter
Real-world testing showed improved accuracy with up to 15.3% reduction in distance estimation errors
The technology has been validated against motion capture ground truth data
📖 Full Retelling
Researchers Luka Šiktar, Branimir Ćaran, Bojan Šekoranja, and Marko Švaco developed a new system for unmanned aerial vehicles (UAVs) to accurately estimate and maintain safe distances from people during search and rescue operations, as detailed in their paper submitted to arXiv on February 24, 2026. The innovation combines Extended Kalman Filter (EKF) algorithms with deep learning techniques to enable UAVs to detect, recognize, and follow individuals while maintaining appropriate distances - a critical safety requirement in search and rescue scenarios. This technology addresses the challenge of precise distance estimation between UAVs and human targets under real-world conditions, which is essential for both operational effectiveness and safety during emergency response missions. The research team's approach represents a significant advancement in autonomous search and rescue capabilities, potentially saving lives by enabling faster and more reliable deployment of UAVs in disaster scenarios where human lives are at risk.
🏷️ Themes
Search and Rescue Technology, UAV Robotics, Computer Vision
In machine learning, deep learning focuses on utilizing multilayered neural networks to perform tasks such as classification, regression, and representation learning. The field takes inspiration from biological neuroscience and revolves around stacking artificial neurons into layers and "training" t...
An unmanned aerial vehicle (UAV) or unmanned aircraft system (UAS), commonly known as a drone, is an aircraft with no human pilot, crew, or passengers on board, but rather is controlled remotely or is autonomous. UAVs were originally developed through the twentieth century for military missions too ...
Search for and provision of aid to people who are in distress or imminent danger
Search and rescue (SAR) is the search for and provision of aid to people who are in distress or imminent danger. The general field of search and rescue includes many specialty sub-fields, typically determined by the type of terrain the search is conducted over. These include mountain rescue; ground ...
--> Computer Science > Robotics arXiv:2602.20958 [Submitted on 24 Feb 2026] Title: EKF-Based Depth Camera and Deep Learning Fusion for UAV-Person Distance Estimation and Following in SAR Operations Authors: Luka Šiktar , Branimir Ćaran , Bojan Šekoranja , Marko Švaco View a PDF of the paper titled EKF-Based Depth Camera and Deep Learning Fusion for UAV-Person Distance Estimation and Following in SAR Operations, by Luka \v iktar and 2 other authors View PDF HTML Abstract: Search and rescue operations require rapid responses to save lives or property. Unmanned Aerial Vehicles equipped with vision-based systems support these missions through prior terrain investigation or real-time assistance during the mission itself. Vision-based UAV frameworks aid human search tasks by detecting and recognizing specific individuals, then tracking and following them while maintaining a safe distance. A key safety requirement for UAV following is the accurate estimation of the distance between camera and target object under real-world conditions, achieved by fusing multiple image modalities. UAVs with deep learning-based vision systems offer a new approach to the planning and execution of SAR operations. As part of the system for automatic people detection and face recognition using deep learning, in this paper we present the fusion of depth camera measurements and monocular camera-to-body distance estimation for robust tracking and following. Deep learning-based filtering of depth camera data and estimation of camera-to-body distance from a monocular camera are achieved with YOLO-pose, enabling real-time fusion of depth information using the Extended Kalman Filter algorithm. The proposed subsystem, designed for use in drones, estimates and measures the distance between the depth camera and the human body keypoints, to maintain the safe distance between the drone and the human target. Our system provides an accurate estimated distance, which has been validated against motion capture g...
