Low-Latency Detection and Tracking of Aircraft in Very High-Resolution Video Feeds
Applying machine learning techniques for real-time detection and tracking of objects in very high-resolution video is a problem that has not been extensively studied. In this thesis, the practical uses of object detection for airport remote towers are explored. We present a Kalman filter-based track...
Main Author: | |
---|---|
Format: | Others |
Language: | English |
Published: |
Linköpings universitet, Interaktiva och kognitiva system
2018
|
Subjects: | |
Online Access: | http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-148848 |
id |
ndltd-UPSALLA1-oai-DiVA.org-liu-148848 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-UPSALLA1-oai-DiVA.org-liu-1488482018-06-26T06:09:44ZLow-Latency Detection and Tracking of Aircraft in Very High-Resolution Video FeedsengLåglatent detektion och spårning av flygplan i högupplösta videokällorMathiesen, JarleLinköpings universitet, Interaktiva och kognitiva system2018trackingobject trackingkalman filterdeep learningremote towerconvolutional neural networkcnnreal-timeComputer SciencesDatavetenskap (datalogi)Applying machine learning techniques for real-time detection and tracking of objects in very high-resolution video is a problem that has not been extensively studied. In this thesis, the practical uses of object detection for airport remote towers are explored. We present a Kalman filter-based tracking framework for low-latency aircraft tracking in very high-resolution video streams. The object detector was trained and tested on a dataset containing 3000 labelled images of aircrafts taken at Swedish airports, reaching an mAP of 90.91% with an average IoU of 89.05% on the test set. The tracker was benchmarked on remote tower video footage from Örnsköldsvik and Sundsvall using slightly modified variants of the MOT-CLEAR and ID metrics for multiple object trackers, obtaining an IDF1 score of 91.9%, and a MOTA score of 83.3%. The prototype runs the tracking pipeline on seven high resolution cameras simultaneously at 10 Hz on a single thread, suggesting large potential speed gains being attainable through parallelization. Student thesisinfo:eu-repo/semantics/bachelorThesistexthttp://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-148848application/pdfinfo:eu-repo/semantics/openAccess |
collection |
NDLTD |
language |
English |
format |
Others
|
sources |
NDLTD |
topic |
tracking object tracking kalman filter deep learning remote tower convolutional neural network cnn real-time Computer Sciences Datavetenskap (datalogi) |
spellingShingle |
tracking object tracking kalman filter deep learning remote tower convolutional neural network cnn real-time Computer Sciences Datavetenskap (datalogi) Mathiesen, Jarle Low-Latency Detection and Tracking of Aircraft in Very High-Resolution Video Feeds |
description |
Applying machine learning techniques for real-time detection and tracking of objects in very high-resolution video is a problem that has not been extensively studied. In this thesis, the practical uses of object detection for airport remote towers are explored. We present a Kalman filter-based tracking framework for low-latency aircraft tracking in very high-resolution video streams. The object detector was trained and tested on a dataset containing 3000 labelled images of aircrafts taken at Swedish airports, reaching an mAP of 90.91% with an average IoU of 89.05% on the test set. The tracker was benchmarked on remote tower video footage from Örnsköldsvik and Sundsvall using slightly modified variants of the MOT-CLEAR and ID metrics for multiple object trackers, obtaining an IDF1 score of 91.9%, and a MOTA score of 83.3%. The prototype runs the tracking pipeline on seven high resolution cameras simultaneously at 10 Hz on a single thread, suggesting large potential speed gains being attainable through parallelization. |
author |
Mathiesen, Jarle |
author_facet |
Mathiesen, Jarle |
author_sort |
Mathiesen, Jarle |
title |
Low-Latency Detection and Tracking of Aircraft in Very High-Resolution Video Feeds |
title_short |
Low-Latency Detection and Tracking of Aircraft in Very High-Resolution Video Feeds |
title_full |
Low-Latency Detection and Tracking of Aircraft in Very High-Resolution Video Feeds |
title_fullStr |
Low-Latency Detection and Tracking of Aircraft in Very High-Resolution Video Feeds |
title_full_unstemmed |
Low-Latency Detection and Tracking of Aircraft in Very High-Resolution Video Feeds |
title_sort |
low-latency detection and tracking of aircraft in very high-resolution video feeds |
publisher |
Linköpings universitet, Interaktiva och kognitiva system |
publishDate |
2018 |
url |
http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-148848 |
work_keys_str_mv |
AT mathiesenjarle lowlatencydetectionandtrackingofaircraftinveryhighresolutionvideofeeds AT mathiesenjarle laglatentdetektionochsparningavflygplanihogupplostavideokallor |
_version_ |
1718707752747925504 |