東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Determination of the Location, Size,...

Dreyer, Matthew,

FindBook

Google Book

Amazon

博客來

Determination of the Location, Size, and Geometry of a Multi-Rotor UAV from Two Video Streams /

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Determination of the Location, Size, and Geometry of a Multi-Rotor UAV from Two Video Streams // Matthew Dreyer.
作者:	Dreyer, Matthew,
面頁冊數:	1 electronic resource (99 pages)
附註:	Source: Masters Abstracts International, Volume: 85-01.
Contained By:	Masters Abstracts International85-01.
標題:	Aerospace engineering. -
電子資源:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30569299
ISBN:	9798379903312

Determination of the Location, Size, and Geometry of a Multi-Rotor UAV from Two Video Streams /
Dreyer, Matthew,

Determination of the Location, Size, and Geometry of a Multi-Rotor UAV from Two Video Streams /Matthew Dreyer. - 1 electronic resource (99 pages)

Source: Masters Abstracts International, Volume: 85-01.

In recent years, drones have solidified as a common tool in a number of fields like farming, package delivery, construction, etc., and with their operations expected to grow, a more drone infested future is eminent. Now, these drones are under surveillance by current FAA regulations, but with more "air traffic", regulations may need to be altered to account for a busier sky. Anticipating such a scenario, methodologies need to be developed to prevent unauthorized drones from encroaching restricted airspace, like an airport. This master's thesis proposes a low-cost C-UAS solution by merely utilizing two video feeds to autonomously estimate characteristics (like GPS, size and geometry) of a drone flying in the field of view (FOV) of two cameras. To determine the 3-dimensional GPS coordinates of the drone, photogrammetry techniques require the 2-dimensional pixel location of the drone in each video feed to triangulate the 3-dimensional coordinates. YOLOv3 with a Darknet 53 backbone was chosen as the machine learning model for this effort, because it is one of the best one-stage detectors for small objects like a drone. By calibrating the cameras and knowing the pixel coordinates of an object (given from the machine learning model) in both views, the two sets of pixel coordinates can be utilized to calculate the 3-dimensional location of the object, the drone in this case. Background subtraction provides a relatively easy way to analyze the pixel area of the drone, and since the pixel location of the drone is obtained via the machine learning model, there is minimal noise. Knowing the pixel area and distance the drone is from a calibrated camera, the drone's pixel dimensions can be converted into a real world length, like feet, inches, etc. Background subtraction also supplements the capability to estimate the number of 'arms' on the multirotor. After converting the drone from RGB to binary with background subtraction, a radial sweep is conducted around the center of the drone, and the "length" of the active pixels at angular intervals are recorded and plotted. By counting the number of peaks in the plot, the number of arms can be estimated for a given frame in a video feed. As a result, the location, geometry, and size logic were meshed together and deployed on a pre-recorded flight of the drone and estimates of each parameter were made online.

English

ISBN: 9798379903312Subjects--Topical Terms:

1002622
Aerospace engineering.
Subjects--Index Terms:

Counter Unmanned Aerial System

Determination of the Location, Size, and Geometry of a Multi-Rotor UAV from Two Video Streams /
LDR:03772nmm a22004213i 4500 001 2438398
005 20260727132404.5
006 m o d
007 cr|nu||||||||
008 261011s2023 miu||||||m |||||||eng d
020 $a 9798379903312
035 $a (MiAaPQD)AAI30569299
035 $a AAI30569299
040 $a MiAaPQD $b eng $c MiAaPQD $e rda
100 1 $a Dreyer, Matthew, $e author. $3 3829688
245 1 0 $a Determination of the Location, Size, and Geometry of a Multi-Rotor UAV from Two Video Streams / $c Matthew Dreyer.
264 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2023
300 $a 1 electronic resource (99 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Masters Abstracts International, Volume: 85-01.
500 $a Advisors: Gururajan, Srikanth Committee members: LaForest, Lisa; Hou, Jie.
502 $b M.S. $c Saint Louis University $d 2023.
520 $a In recent years, drones have solidified as a common tool in a number of fields like farming, package delivery, construction, etc., and with their operations expected to grow, a more drone infested future is eminent. Now, these drones are under surveillance by current FAA regulations, but with more "air traffic", regulations may need to be altered to account for a busier sky. Anticipating such a scenario, methodologies need to be developed to prevent unauthorized drones from encroaching restricted airspace, like an airport. This master's thesis proposes a low-cost C-UAS solution by merely utilizing two video feeds to autonomously estimate characteristics (like GPS, size and geometry) of a drone flying in the field of view (FOV) of two cameras. To determine the 3-dimensional GPS coordinates of the drone, photogrammetry techniques require the 2-dimensional pixel location of the drone in each video feed to triangulate the 3-dimensional coordinates. YOLOv3 with a Darknet 53 backbone was chosen as the machine learning model for this effort, because it is one of the best one-stage detectors for small objects like a drone. By calibrating the cameras and knowing the pixel coordinates of an object (given from the machine learning model) in both views, the two sets of pixel coordinates can be utilized to calculate the 3-dimensional location of the object, the drone in this case. Background subtraction provides a relatively easy way to analyze the pixel area of the drone, and since the pixel location of the drone is obtained via the machine learning model, there is minimal noise. Knowing the pixel area and distance the drone is from a calibrated camera, the drone's pixel dimensions can be converted into a real world length, like feet, inches, etc. Background subtraction also supplements the capability to estimate the number of 'arms' on the multirotor. After converting the drone from RGB to binary with background subtraction, a radial sweep is conducted around the center of the drone, and the "length" of the active pixels at angular intervals are recorded and plotted. By counting the number of peaks in the plot, the number of arms can be estimated for a given frame in a video feed. As a result, the location, geometry, and size logic were meshed together and deployed on a pre-recorded flight of the drone and estimates of each parameter were made online.
546 $a English
590 $a School code: 0193
650 4 $a Aerospace engineering. $3 1002622
650 4 $a Computer science. $3 523869
650 4 $a Geographic information science. $3 3432445
653 $a Counter Unmanned Aerial System
653 $a Computer vision
653 $a Drones
653 $a Pursuit evasion
690 $a 0538
690 $a 0984
690 $a 0370
710 2 $a Saint Louis University. $b Engineering. $e degree granting institution. $3 3829689
720 1 $a Gururajan, Srikanth $e degree supervisor.
773 0 $t Masters Abstracts International $g 85-01.
790 $a 0193
791 $a M.S.
792 $a 2023
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30569299