Open Access Repository

Multi-sensor, multi-temporal, and ultra-high resolution environmental remote sensing from UAVs

Downloads

Downloads per month over past year

Turner, D (2015) Multi-sensor, multi-temporal, and ultra-high resolution environmental remote sensing from UAVs. PhD thesis, University of Tasmania.

[img]
Preview
PDF (Whole thesis)
Turner_whole_th...pdf | Download (7MB)
Available under University of Tasmania Standard License.

| Preview

Abstract

Civilian use of Unmanned Aerial Vehicles (UAVs) has become increasingly common in recent times. Improvements in airframe design and electronics, particularly the mass production of comparatively inexpensive miniaturised inertial and positioning sensors, has enabled the application of UAVs to many and varied tasks. One area of growth in the scientific community has been the use of UAVs for Environmental Remote Sensing (ERS) where high spatial and temporal resolution, the ability to fly on-demand, and data collection from multiple sensors offers substantial advantages over traditional techniques.
For small regions, Micro-UAVs (MUAVs), typically weighing less than 5 kg with flying duration of up to 30 minutes, present an excellent option for collecting the required remotely sensed data for understanding environmental processes that require high resolution (centimetre scale), multi-sensor data. There are, however, several important issues to be considered and further developed. The high resolution images have a small footprint and in most cases, hundreds of overlapping images are required to span the region of interest. These images often contain high perspective distortions (compared to traditional nadir aerial photography) and thus can be difficult to process with conventional techniques and software. For most applications, it is important that the imagery is accurately georeferenced, which is typically undertaken using Ground Control Points (GCPs). Collection of GCPs can be a time-consuming process and detracts from one of the advantages of an MUAV, which is operational flexibility and efficiency. In addition, to acquire multi-sensor datasets, an MUAV will need to carry each of the multiple sensors on separate flights, which means the image datasets from each of these flights need to be accurately co-registered. Finally, if repeat coverage is required over time, data collection and processing methods must be robust and repeatable.
This thesis sets out to address these barriers, particularly those associated with processing high resolution imagery collected with multiple sensors. The broad aim of this study is to determine appropriate workflows to enable the efficient, timely, and accurate processing of multi-sensor data collected from an MUAV. Case studies are used to demonstrate how specific challenges are addressed and to quantify the accuracy achieved in the context of various environmental monitoring applications.
A methodology to geometrically correct and mosaic UAV imagery using feature matching and Structure from Motion (SfM) photogrammetric techniques was developed. This technique is fully automated and can georectify and mosaic imagery based either on GCPs (achieving an accuracy of 10 – 15 cm) or via a Direct Georeferencing (DG) technique (with an accuracy of 65–120 cm when using the navigation-grade on-board GPS). The DG system, which used the location of the camera at time of exposure as the basis for georeferencing, was limited by the accuracy of the GPS used to measure airframe position (generally a navigation-grade receiver) and the accuracy of the synchronisation between time of exposure and the GPS position record. A camera-GPS module was developed that incorporated a higher accuracy GPS (single frequency carrier phase based unit with an accuracy of 10 – 20 cm) and a camera synchronisation system. Commercial software was used to process and directly georeference the imagery and achieve an absolute spatial accuracy of 11 cm, which is commensurate with the accuracy of the GPS unit used.
A case study that investigated the physiological state of Antarctic moss ecosystems was used to demonstrate that data from multiple sensors can be accurately co-registered. Imagery from each sensor was georeferenced and mosaicked with a combination of commercially available software and custom routines that were based on the Scale Invariant Feature Transform (SIFT) and SfM workflow. The spatial co-registration of the mosaics was measured and found to have a mean root mean squared error (RMSE) of 1.78 pixels. This study also demonstrated that quantitative data can be collected with specialised sensors and then related to plant traits. In particular, the Modified Triangular Vegetation Index (MTVI) was derived from the multispectral data and related to the health of moss quadrats (as measured in-situ) and a statistically significant (R2 = 0.64) relationship was found.
The ability of MUAVs to be used for time series analysis was demonstrated with a case study of a highly dynamic landslide that was monitored from 2010 through to 2014 with seven separate datasets collected during the period. Software based on SfM algorithms was used to create Digital Surface Models (DSMs) of the landslide surface with an accuracy of around 4 – 5 cm in the horizontal and 3 – 4 cm in the vertical. The accuracy of the co-registration of subsequent DSMs was checked and corrected based on comparing non-active areas of the landslide, which minimised alignment errors to a mean of 7 cm. It was discovered that the methodology could also be applied to historical aerial photography to create a baseline DSM allowing the total displacement of the landslide to be calculated (approximately 6630 m3). This study demonstrated that MUAVs can be used repeatedly to map the dynamics of a landslide over a period of 4 years.
Addressing the issues presented throughout this thesis demonstrates the clear potential of MUAVs for a wide range of applications within the broad discipline of Environmental Remote Sensing. It was also shown that MUAVs offer a series of benefits such as high spatial and temporal resolutions along with the ability to collect multi-sensor data. Ongoing technological developments, particularly in sensor miniaturisation, high capacity power storage, autopilot reliability, and motor design will likely continue the present upward trajectory of MUAV use across the diverse user communities.

Item Type: Thesis (PhD)
Keywords: UAV, image processing, environmental remote sensing
Copyright Holders: Copyright 2015 the author
Additional Information:

Chapter 2 has been published under a Creative Commons License as: Turner, D., Lucieer, A., Watson C. (2012), An Automated technique for generating georectified mosaics from ultra-high resolution unmanned aerial vehicle (UAV) imagery, based on structure from motion (SfM) point clouds, Remote sensing 4(5), 1392-1410.

Chapter 3 appears to be the equivalent of a post-print version of an article published as: Turner, D., Lucieer, A., Wallace, (2013), Direct georeferencing of ultrahigh-resolution UAV imagery, IEEE transactions on geoscience and remote sensing, 52(5), 2738-2745 © © 2013 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

Chapter 4 has been published under a Creative Commons License as: Turner, D., Lucieer, A. Malenovský, Z. King, D., Robinson, S. (2014), Spatial co-registration of ultra-high resolution visible, multispectral and thermal images acquired with a micro-UAV over Antarctic moss beds, Remote sensing 6(5, 4003-4024.

Chapter 5 has been published under a Creative Commons License as: Turner, D., Lucieer, A., de Jong, S. M., (2015), Time series analysis of landslide dynamics using an unmanned aerial vehicle (UAV), Remote sensing, 7 (2), 1736-1757.

Date Deposited: 01 Sep 2016 02:08
Last Modified: 01 Sep 2016 02:08
Related URLs:
Item Statistics: View statistics for this item

Actions (login required)

Item Control Page Item Control Page
TOP