Development Of Surface Surveillance System Using Autonomous Drones
The development of drones is significant in the past few years. Drones are more reliable and common now. People started to realise their potential and used it in the surveillance industry. Universiti Tunku Abdul Rahman (UTAR) Kampar had decided to use drones for the surveillance land owned by UTAR t...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Year Project / Dissertation / Thesis |
| Published: |
2021
|
| Subjects: | |
| Online Access: | http://eprints.utar.edu.my/4215/1/1605635_FYP_Report_%2D_WEN_QING_LIM.pdf http://eprints.utar.edu.my/4215/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The development of drones is significant in the past few years. Drones are more reliable and common now. People started to realise their potential and used it in the surveillance industry. Universiti Tunku Abdul Rahman (UTAR) Kampar had decided to use drones for the surveillance land owned by UTAR to prevent unauthorised construction and plantation. “DJI Mavic 2 Enterprise Dual” is selected and responsible for the mapping mission via DJI Pilot. A total of 1741 images are captured to reconstruct the two-dimensional (2D) and threedimensional (3D) map of the rural area which is owned by UTAR. An automated image sorting system is created to sort the images due to their large size and repeatability. The reconstruction of images is carried out by WebODM. The custom parameters for the reconstruction are dem-resolution: 15, ignoregsd: true, min-num-features: 20000, orthophoto-resolution: 15, skip-3dmodel: true, texturing-data-term: area. It is specifically built for the reconstruction of the area, which is full of trees and green vegetation. 2D map is a digital orthophoto map (DOM). These images covered an area of around 1 971 000 Sq Meters (488 Acres). After obtaining the 2D map, the 2D map in different timestamps will be used to compare and find out the difference to achieve the objective of surface surveillance. This process is known as change detection. The change detection is done by image registration followed by principal component analysis (PCA) and K-mean clustering on the “difference image” to cluster out changes and non-changes. A cluster of four is used for the K-mean clustering to minimise the noises and false-positive detection. |
|---|