Advancements in Aerial Surveillance and Object Detection

The field of aerial surveillance and object detection is rapidly advancing with the development of new technologies and techniques. Researchers are exploring new approaches to improve the accuracy and efficiency of object detection in aerial images and videos, including the use of deep learning methods, multi-modal fusion, and meta-learning. One of the key challenges in this field is the ability to detect and track small objects in complex environments, such as low-altitude aircraft or maritime objects. To address this challenge, researchers are proposing new datasets and benchmarking protocols to evaluate the performance of object detection algorithms in these scenarios. Noteworthy papers in this area include SatSAM2, which proposes a zero-shot satellite video tracker that outperforms state-of-the-art methods, and LAA3D, which introduces a large-scale dataset for 3D detection and tracking of low-altitude aerial vehicles. Additionally, papers like MambaRefine-YOLO and IrisNet are proposing new methods for small object detection in UAV imagery, leveraging techniques such as dual-modality fusion and meta-learning. Overall, the field of aerial surveillance and object detection is seeing significant advancements, with a focus on improving the accuracy and efficiency of object detection algorithms in complex environments.

Sources

Person Recognition in Aerial Surveillance: A Decade Survey

SatSAM2: Motion-Constrained Video Object Tracking in Satellite Imagery using Promptable SAM2 and Kalman Priors

A Tri-Modal Dataset and a Baseline System for Tracking Unmanned Aerial Vehicles

AIRHILT: A Human-in-the-Loop Testbed for Multimodal Conflict Detection in Aviation

LAA3D: A Benchmark of Detecting and Tracking Low-Altitude Aircraft in 3D Space

MambaRefine-YOLO: A Dual-Modality Small Object Detector for UAV Imagery

Dual-Granularity Semantic Prompting for Language Guidance Infrared Small Target Detection

Maritime Small Object Detection from UAVs using Deep Learning with Altitude-Aware Dynamic Tiling

IrisNet: Infrared Image Status Awareness Meta Decoder for Infrared Small Targets Detection

MODEST: Multi-Optics Depth-of-Field Stereo Dataset

AerialMind: Towards Referring Multi-Object Tracking in UAV Scenarios

UAVLight: A Benchmark for Illumination-Robust 3D Reconstruction in Unmanned Aerial Vehicle (UAV) Scenes

Built with on top of