Robust Target Detection snf Tracking in Unmanned Aerial Vehicles (UAVs) Imagery (Record no. 8808)

MARC details
000 -LEADER
fixed length control field 10041nam a22002537a 4500
008 - FIXED-LENGTH DATA ELEMENTS--GENERAL INFORMATION
fixed length control field 210112b2013 a|||f mb|| 00| 0 eng d
040 ## - CATALOGING SOURCE
Original cataloging agency EG-CaNU
Transcribing agency EG-CaNU
041 0# - Language Code
Language code of text eng
Language code of abstract eng
082 ## - DEWEY DECIMAL CLASSIFICATION NUMBER
Classification number 610
100 0# - MAIN ENTRY--PERSONAL NAME
Personal name Mennatullah Mohamed Siam
245 1# - TITLE STATEMENT
Title Robust Target Detection snf Tracking in Unmanned Aerial Vehicles (UAVs) Imagery
Statement of responsibility, etc. Mennatullah Mohamed Siam
260 ## - PUBLICATION, DISTRIBUTION, ETC.
Date of publication, distribution, etc. 2013
300 ## - PHYSICAL DESCRIPTION
Extent p.
Other physical details ill.
Dimensions 21 cm.
500 ## - GENERAL NOTE
Materials specified Supervisor: Mohamed A. El-Helw
502 ## - Dissertation Note
Dissertation type Thesis (M.A.)—Nile University, Egypt, 2013 .
504 ## - Bibliography
Bibliography "Includes bibliographical references"
505 0# - Contents
Formatted contents note Contents:<br/>Chapter1: Introduction ........................................................................................................ 1<br/>1.1 Overview and Motivation ............................................................................... 1<br/>1.2 Problem Definition ......................................................................................... 3<br/>1.3 Contributions .................................................................................................. 3<br/>1.3.1 Real-time Onboard Target Detection and Tracking ................................ 3<br/>1.3.2 Robust Tracking using P-N Learning ..................................................... 4<br/>1.4 Publications .................................................................................................... 5<br/>1.5 Organization of Thesis ................................................................................... 6<br/>Chapter 2: Review of Detection and Tracking in UAV Imagery ....................................... 7<br/>2.1 Unmanned Aerial Vehicles (UAVs) .............................................................. 7<br/>2.1.1 UAVs Types............................................................................................ 7<br/>2.1.2 Existing UAV Models............................................................................. 8<br/>2.1.3 UAVs Applications ............................................................................... 10<br/>2.2 Target Detection and Tracking ..................................................................... 11<br/>2.2.1 Target Detection Taxonomy ................................................................. 11<br/>2.2.2 Target Tracking Taxonomy .................................................................. 13<br/>2.3 Target Detection in UAV Imagery ............................................................... 15<br/>2.3.1 Interest Point Detectors ......................................................................... 15<br/>2.3.2 Motion Detection .................................................................................. 16<br/>2.3.3 Segmentation......................................................................................... 16<br/>2.3.4 Supervised Learning Detection ............................................................. 17<br/>vii<br/>2.3.5 Summary ............................................................................................... 18<br/>2.4 Target Tracking in UAV Imagery ................................................................ 19<br/>2.4.1 Point Tracking ....................................................................................... 19<br/>2.4.2 Kernel Tracking .................................................................................... 20<br/>2.5 State of the Art Trackers .............................................................................. 21<br/>Chapter 3: Integrated Framework for Target Detection and Tracking ............................. 23<br/>3.1 Framework Architecture .............................................................................. 23<br/>3.2 Multiple Target Detection ............................................................................ 24<br/>3.2.1 Feature Extractor ................................................................................... 24<br/>3.2.2 Motion Estimation ................................................................................ 25<br/>3.2.3 Outliers Detection and DBScan Clustering .......................................... 30<br/>3.3. Target Tracking ............................................................................................ 34<br/>3.3.1. Overlap-Rate Based Data Association .................................................. 34<br/>3.3.2. Kalman Prediction ................................................................................ 37<br/>3.3.3. BRIEF Descriptors Search .................................................................... 39<br/>Chapter 4: Enhanced Target Tracking using PN Constraints ........................................... 42<br/>4.1 Enhanced Version Framework Architecture ................................................ 42<br/>4.2 Updated Data Association Algorithm .......................................................... 43<br/>4.3 P-N (Positive-Negative) Learning ................................................................ 44<br/>4.3.1 Semi-supervised Learning Overview .................................................... 44<br/>4.3.2 P-N Learning Approach ........................................................................ 45<br/>4.3.3 P-N Learning in Object Tracking ......................................................... 46<br/>4.4 Novel P-N Experts ....................................................................................... 49<br/>Chapter 5: Results & Discussion ...................................................................................... 51<br/>5.1 Datasets ........................................................................................................ 51<br/>5.1.1. Quad-rotor ............................................................................................. 52<br/>5.1.2. PandaBoard and Middleware ................................................................ 53<br/>5.1.3. Dataset Collection ................................................................................. 54<br/>5.2 Evaluation Test-bed and Metrics .................................................................. 55<br/>5.2.1 Evaluation Test-bed .............................................................................. 55<br/>5.2.2 Evaluation Metrics ................................................................................ 56<br/>viii<br/>5.3 Qualitative Analysis ..................................................................................... 57<br/>5.3.1 Recapture Targets Out/In Camera Field of View: ................................ 61<br/>5.4 Quantitative Evaluation ................................................................................ 63<br/>Chapter 6: Conclusions and Future Work ......................................................................... 69<br/>6.1. Summary of Contributions ........................................................................... 69<br/>6.2. Future Work .................................................................................................
520 3# - Abstract
Abstract Abstract:<br/>In this dissertation we address the problem of automatic target detection and<br/>tracking in UAV imagery. Even though target detection and tracking are well<br/>explored problems in the field of computer vision, several challenges arise in case of<br/>UAV imagery due to the modus operandi of the latter in terms of operational and<br/>processing constraints. UAVs are airborne platforms that are typically in continuous<br/>and sometimes turbulent rapid motions. The payloads are minimal with limited<br/>processing boards and low-resolution image sensors capturing imagery from high<br/>altitudes. Consequently, on-board vision-based algorithms have to decouple scene<br/>and UAV motions to detect targets of interest, shrewdly exploit low-resolution<br/>imagery to discriminate targets based on appearance, and execute efficiently to meet<br/>the necessary demand of real-time performance. Additional challenges include the<br/>fact that tracking of detected targets should be adaptive and learns the appearance of<br/>the targets to maintain their tracking case of significant illumination changes and<br/>partial occlusions, re-acquire targets as they move out of and subsequently in the<br/>field-of-view, and to discriminate between similar targets moving within close spatial<br/>proximity of each other.<br/>This thesis presents an integrated framework comprising novel set of algorithms<br/>to alleviate the above described problems in order to facilitate on-board robust realtime<br/>target detection and tracking in UAV imagery. The separation of scene and<br/>camera motions from the targets' motion is achieved by using an algorithm based on<br/>image feature processing and projective geometry. Detected targets are tracked with<br/>Kalman filtering while an overlap-rate-based data association mechanism followed by<br/>tracking persistency check are used to discriminate between true moving targets and<br/>false detections. The proposed framework doesn’t involve explicit application of<br/>image transformations to detect potential targets resulting in enhanced computational<br/>time and reduction of registration errors. The tracking performance is further<br/>enhanced by utilizing P-N learning to learn the targets' appearances online. To this<br/>end, novel P-N constraints are introduced based on data association to control the<br/>xiii<br/>positive and negative samples while a cascaded classifier is employed to detect the<br/>targets in case of association failure after learning targets' appearance. Experimental<br/>results obtained with publicly available DARPA aerial datasets demonstrate that the<br/>proposed tracker with automatic detection feedback achieves better recall and average<br/>overlap than existing manually-initialized trackers. Further qualitative analysis proves<br/>the framework is able to solve most of the above challenges including target recapture<br/>and discriminating between targets with similar appearance while alleviating tracker<br/>drift problems.
546 ## - Language Note
Language Note Text in English, abstracts in English.
650 #4 - Subject
Subject Informatics-IFM
655 #7 - Index Term-Genre/Form
Source of term NULIB
focus term Dissertation, Academic
690 ## - Subject
School Informatics-IFM
942 ## - ADDED ENTRY ELEMENTS (KOHA)
Source of classification or shelving scheme Dewey Decimal Classification
Koha item type Thesis
650 #4 - Subject
-- 266
655 #7 - Index Term-Genre/Form
-- 187
690 ## - Subject
-- 266
Holdings
Withdrawn status Lost status Source of classification or shelving scheme Damaged status Not for loan Home library Current library Date acquired Total Checkouts Full call number Date last seen Price effective from Koha item type
    Dewey Decimal Classification   Not For Loan Main library Main library 01/12/2021   610/ MS.R 2013 01/12/2021 01/12/2021 Thesis