Small target detection for search and rescue operations using distributed deep learning and synthetic data generation

  title={Small target detection for search and rescue operations using distributed deep learning and synthetic data generation},
  author={Kyongsik Yun and Luan Nguyen and Tuan T. Nguyen and Doyoung Kim and Sarah Eldin and Alexander Huyen and Thomas T. Lu and Edward Chow},
  booktitle={Defense + Commercial Sensing},
It is important to find the target as soon as possible for search and rescue operations. Surveillance camera systems and unmanned aerial vehicles (UAVs) are used to support search and rescue. Automatic object detection is important because a person cannot monitor multiple surveillance screens simultaneously for 24 hours. Also, the object is often too small to be recognized by the human eye on the surveillance screen. This study used an UAVs around the Port of Houston and fixed surveillance… 

Automatic Person Detection in Search and Rescue Operations Using Deep CNN Detectors

The paper proposes a model that can be used in SAR operations because of the excellent results in detecting people in search and rescue scenarios, and proposes the YOLOv4 detector, chosen for further examination.

Deep Learning-based Human Detection for UAVs with Optical and Infrared Cameras: System and Experiments

A deep learning-based human detection system that uses optical (RGB) and long-wave infrared (LWIR) cameras to detect, track, localize, and re-identify humans from UAVs flying at high altitude shows that by optimizing the bounding box anchors and augmenting the image resolution the number of missed detections from high altitudes can be decreased by over 20 percent.

Person Detection in Drone Imagery

A significant improvement in the detection results of persons in the images taken by the drone was achieved concerning mAP and precision and recall and on a custom SARD image dataset built to simulate search and rescue scenes.

Multimodel Deep Learning for Person Detection in Aerial Images

A novel method for person detection in aerial images of nonurban terrain gathered by an UAV, which plays an important role in Search And Rescue (SAR) missions, is proposed and is suitable for usage as an auxiliary method in real SAR operations.

SaRNet: A Dataset for Deep Learning Assisted Search and Rescue with Satellite Imagery

  • Michael ThoreauFrazer Wilson
  • Environmental Science, Computer Science
    2021 12th International Symposium on Image and Signal Processing and Analysis (ISPA)
  • 2021
A novel remote sensing object detection dataset for deep learning assisted SaR, which contains only small objects that have been identified as potential targets as part of a live SaR response and proposes a novel object detection metric, specifically designed to be used in a deep learning assist SaR setting.

Multiple Infrared Small Targets Detection Based on Hierarchical Maximal Entropy Random Walk

A multiple small targets detection method derived from hierarchical maximal entropy random walk (HMERW) based on a proposed graph decomposition theory that is superior to the state-of-the-art methods in terms of target enhancement, background suppression, and multiple small target detection.

The Use of UAVs to Save People in the Mountains: Tasks, Problems, Perspective

A predominant approach was drawn, which allows a high degree of certainty to detect problems of UAV icing and timely change of control system settings, which gives rise to an uncontrolled change in the properties of the control object UAVs.

TruePAL: an AI assistant for first responder safety

The preliminary test simulation results show that TruePAL could provide the drivers and crew members advanced warning before a crash occurs, and the team has successfully demonstrated its capability of timely warning in over a dozen scenarios.

Dynamic fire and smoke detection and classification for flashover prediction

Deep neural networks (DNN) have been used to detect, classify and track fire and smoke areas and aFire and smoke progression curve has been drawn to predict the flashover point.

Synthesizing Retro Game Screenshot Datasets for Sprite Detection

An object detection model from deep learning literature is re-purposed, developing along the way YOLO Artificial Retro-game Data Synthesizer, or YARDS, which efficiently produces semirealistic, retro-game sprite detection datasets without manual labeling.



Analysis of saliency object detection algorithms for search and rescue operations

The possibilities of using saliency object detection algorithms on the problem of detection of objects for Search and Rescue operations (SAR) in UAV images are explored.

A UAV Search and Rescue Scenario with Human Body Detection and Geolocalization

This paper presents a challenging emergency services mission scenarios which involves search and rescue for injured civilians by UAVs and shows how far the UAVTech Lab has come in implementing and executing such a challenging mission in realistic urban scenarios.

Detecting Small Signs from Large Images

Experimental results on a real-world conditioned traffic sign dataset have demonstrated the effectiveness of the proposed approach for detecting small traffic signs from large images under real world conditions in terms of detection accuracy and recall, especially for those with small sizes.

Towards using Unmanned Aerial Vehicles (UAVs) in Wilderness Search and Rescue: Lessons from field trials

Surprisingly to us, sophisticated autonomous search patterns were less critical than the authors anticipated, though advances in video enhancement and visualizing search progress, as well as ongoing work to model the likely location of a missing person, open up the possibility of closing the loop between UAV path-planning, search quality, and the likely locations of a moving missing person.

Feature-fused SSD: fast detection for small objects

This paper proposes a multi-level feature fusion method for introducing contextual information in SSD, in order to improve the accuracy for small objects and designs two feature fusion modules, concatenation module and element-sum module, different in the way of adding contextual information.

The EU-ICARUS project: Developing assistive robotic tools for search and rescue operations

The ICARUS EU-FP7 project deals with the development of a set of integrated components to assist search and rescue teams in dealing with the difficult and dangerous, but life-saving task of finding

Multipurpose UAV for search and rescue operations in mountain avalanche events

A multipurpose UAV (unmanned aerial vehicle) for mountain rescue operations designed to meet environmental requirements for mountainous terrain, assuring the capability of carrying different payloads such as avalanche beacon (ARTVA) with automatic signal recognition and path following algorithms.

MoCap-guided Data Augmentation for 3D Pose Estimation in the Wild

This paper introduces an image-based synthesis engine that artificially augments a dataset of real images with 2D human pose annotations using 3D Motion Capture (MoCap) data to generate a large set of photorealistic synthetic images of humans with 3D pose annotations.

Design of an Autonomous Robot for Mapping, Navigation, and Manipulation in Underground Mines

Analyzes requirements and presents a robot design capable of navigating autonomously underground and manipulating objects with a robotic arm and shows the robot's operability underground.

UnrealCV: Connecting Computer Vision to Unreal Engine

An open-source plugin UnrealCV is created for a popular game engine Unreal Engine 4 (UE4) to enable researchers to build on these resources to create virtual worlds, provided they can access and modify the internal data structures of the games.