Infrastructure Engineering - Theses
Permanent URI for this collection
Search Results
1 - 2 of 2
-
ItemCrop monitoring with Unmanned Aerial Vehicles (UAV)( 2015)Remote sensing is one of the tools used in precision agriculture to map the spatial variability of crops at critical grow stages. Remote sensing using Unmanned Aerial Vehicles (UAV) enables the collection of high resolution imagery to aid in the precision agriculture decision management process and has demonstrated the detection of crop water stress, crop yield estimation, phenotyping and disease detection. However, the commercial adoption of UAV in precision agriculture to date has been limited due to a number of factors including technical and data interpretation challenges, UAV regulations and economic factors. To aid in the adoption and ease of use of UAV for Precision Agriculture, this work addresses two fundamental technical challenges i) reliable remote sensing image acquisition and ii) automated image processing. Accurate forward and side overlap between images ensures robust geometry is captured for image feature matching and stitching. To ensure reliable image acquisition, we implemented an adaptive intervalometer UAV photogrammetric payload system that adjusts image acquisition in real-time to accommodate ground speed and altitude variations. The adaptive intervalometer, evaluated on a multirotor and fixed-wing platform, achieved an image forward overlap RMSE of 0.84% and 2.26% respectively, which is a significant improvement in accuracy compared to the commonly used time-based intervalometer. To illustrate the effectiveness of the UAV payload, the adaptive intervalometer was used to simultaneously collect high resolution thermal and multispectral imagery to detect crop water stress status in an orchard. Several studies have demonstrated that high-resolution visual/near-infrared (VNIR) vineyard maps can be used as a PA tool for multi-temporal monitoring of vineyard spatial variability, shape and vigor to aid in the application of variable-rate treatments and irrigation scheduling. Generating vineyard maps requires separating vine pixels from non-vine pixels in order to accurately determine vine spectral and spatial information. However, manual segmentation of high resolution aerial images is time consuming and costly. Previously, several image texture and frequency analysis methods have been applied to vineyard map generation; however these approaches require manual preliminary delineation of the vine fields. To assist in the automation of vine row classification, we developed an algorithm that uses skeletonisation techniques to reduce the complexity of agricultural scenes into a collection of skeletal descriptors. The algorithm was evaluated on a high resolution aerial orthomosaic and proved its efficiency in unsupervised detection and delineation of vine rows (precision = 0.97) in a commercial vineyard. UAV present new and exciting remote sensing possibilities for environmental monitoring and agricultural applications. By simplify the skill set required to collect and analyse remotely sensed data, we can start to realise the potential UAV have in precision agriculture and environmental monitoring. Our research focuses on horticulture and viticulture UAV applications; however our image acquisition and automated segmentation techniques could be applied to a wide variety of areas including search and rescue, infrastructure management and archeology.
-
ItemThe channels of Mars and the Victorian basalt plains: an investigation of channel chronology( 2012)Current models of Martian channel formation postulate a sudden change in the environment during the Hesperian period. To investigate the chronology of this change, Martian channels with extensive dendritic valley networks where active volcanism was present during the late Noachian to the early Hesperian period of Mars were chosen. Using stereo-derived digital terrain model (DTM) data from the High Resolution Stereo Camera (HRSC) on the orbiter, selected ancient drainage channels on Mars were identified and compared with those on the basaltic plains in Western Victoria, identified on Victorian government digital elevation data. The elevation data from Mars and Victoria were used for stream analysis and to produce longitudinal profiles of streams. An analysis of the maturity of the Victorian stream systems was undertaken using the DEM and the concavity indices of the streams was also calculated. Observation on the maturity of selected stream systems in Victoria was undertaken and the respective sites geologically dated using K-Ar dating of the underlying basaltic plains. The Victorian stream systems are relatively immature by normal terrestrial standards having been formed after flood basalts covered the prior drainage systems. The same stream analysis was then applied to the Martian channels, where there is limited dating control. To better understand the chronology of their formation, the Martian channel concavity indices were calculated and compared to the terrestrial results. Results reveal streams dating up to 1.4 million years ago, narrowing down the time when channel formation ceased to approximately the same amount of time (1.4 Ma) after the volcanic plains were formed in the selected Martian regions.