3D-SiamMask: Vision-Based Multi-Rotor Aerial-Vehicle Tracking for a Moving Object
This paper aims to develop a multi-rotor-based visual tracker for a specified moving object. Visual object-tracking algorithms for multi-rotors are challenging due to multiple issues such as occlusion, quick camera motion, and out-of-view scenarios. Hence, algorithmic changes are required for dealin...
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2022-11-01
|
Series: | Remote Sensing |
Subjects: | |
Online Access: | https://www.mdpi.com/2072-4292/14/22/5756 |
_version_ | 1797464078272167936 |
---|---|
author | Mohamad Al Mdfaa Geesara Kulathunga Alexandr Klimchik |
author_facet | Mohamad Al Mdfaa Geesara Kulathunga Alexandr Klimchik |
author_sort | Mohamad Al Mdfaa |
collection | DOAJ |
description | This paper aims to develop a multi-rotor-based visual tracker for a specified moving object. Visual object-tracking algorithms for multi-rotors are challenging due to multiple issues such as occlusion, quick camera motion, and out-of-view scenarios. Hence, algorithmic changes are required for dealing with images or video sequences obtained by multi-rotors. Therefore, we propose two approaches: a generic object tracker and a class-specific tracker. Both tracking settings require the object bounding box to be selected in the first frame. As part of the later steps, the object tracker uses the updated template set and the calibrated RGBD sensor data as inputs to track the target object using a Siamese network and a machine-learning model for depth estimation. The class-specific tracker is quite similar to the generic object tracker but has an additional auxiliary object classifier. The experimental study and validation were carried out in a robot simulation environment. The simulation environment was designed to serve multiple case scenarios using Gazebo. According to the experiment results, the class-specific object tracker performed better than the generic object tracker in terms of stability and accuracy. Experiments show that the proposed generic tracker achieves promising results on three challenging datasets. Our tracker runs at approximately 36 fps on GPU. |
first_indexed | 2024-03-09T18:01:47Z |
format | Article |
id | doaj.art-33500fca3c8d45f5a1e26e284dbd95fb |
institution | Directory Open Access Journal |
issn | 2072-4292 |
language | English |
last_indexed | 2024-03-09T18:01:47Z |
publishDate | 2022-11-01 |
publisher | MDPI AG |
record_format | Article |
series | Remote Sensing |
spelling | doaj.art-33500fca3c8d45f5a1e26e284dbd95fb2023-11-24T09:49:57ZengMDPI AGRemote Sensing2072-42922022-11-011422575610.3390/rs142257563D-SiamMask: Vision-Based Multi-Rotor Aerial-Vehicle Tracking for a Moving ObjectMohamad Al Mdfaa0Geesara Kulathunga1Alexandr Klimchik2Center for Technologies in Robotics and Mechatronics Components, Innopolis University, 420500 Innopolis, RussiaCenter for Technologies in Robotics and Mechatronics Components, Innopolis University, 420500 Innopolis, RussiaSchool of Computer Science, University of Lincoln, Lincoln LN6 7DL, UKThis paper aims to develop a multi-rotor-based visual tracker for a specified moving object. Visual object-tracking algorithms for multi-rotors are challenging due to multiple issues such as occlusion, quick camera motion, and out-of-view scenarios. Hence, algorithmic changes are required for dealing with images or video sequences obtained by multi-rotors. Therefore, we propose two approaches: a generic object tracker and a class-specific tracker. Both tracking settings require the object bounding box to be selected in the first frame. As part of the later steps, the object tracker uses the updated template set and the calibrated RGBD sensor data as inputs to track the target object using a Siamese network and a machine-learning model for depth estimation. The class-specific tracker is quite similar to the generic object tracker but has an additional auxiliary object classifier. The experimental study and validation were carried out in a robot simulation environment. The simulation environment was designed to serve multiple case scenarios using Gazebo. According to the experiment results, the class-specific object tracker performed better than the generic object tracker in terms of stability and accuracy. Experiments show that the proposed generic tracker achieves promising results on three challenging datasets. Our tracker runs at approximately 36 fps on GPU.https://www.mdpi.com/2072-4292/14/22/5756visual odometrysingle-object trackingdeep learningroboticsunmanned aerial vehicleshigh-accuracy positioning |
spellingShingle | Mohamad Al Mdfaa Geesara Kulathunga Alexandr Klimchik 3D-SiamMask: Vision-Based Multi-Rotor Aerial-Vehicle Tracking for a Moving Object Remote Sensing visual odometry single-object tracking deep learning robotics unmanned aerial vehicles high-accuracy positioning |
title | 3D-SiamMask: Vision-Based Multi-Rotor Aerial-Vehicle Tracking for a Moving Object |
title_full | 3D-SiamMask: Vision-Based Multi-Rotor Aerial-Vehicle Tracking for a Moving Object |
title_fullStr | 3D-SiamMask: Vision-Based Multi-Rotor Aerial-Vehicle Tracking for a Moving Object |
title_full_unstemmed | 3D-SiamMask: Vision-Based Multi-Rotor Aerial-Vehicle Tracking for a Moving Object |
title_short | 3D-SiamMask: Vision-Based Multi-Rotor Aerial-Vehicle Tracking for a Moving Object |
title_sort | 3d siammask vision based multi rotor aerial vehicle tracking for a moving object |
topic | visual odometry single-object tracking deep learning robotics unmanned aerial vehicles high-accuracy positioning |
url | https://www.mdpi.com/2072-4292/14/22/5756 |
work_keys_str_mv | AT mohamadalmdfaa 3dsiammaskvisionbasedmultirotoraerialvehicletrackingforamovingobject AT geesarakulathunga 3dsiammaskvisionbasedmultirotoraerialvehicletrackingforamovingobject AT alexandrklimchik 3dsiammaskvisionbasedmultirotoraerialvehicletrackingforamovingobject |