Unmanned aerial vehicles (UAVs) can provide observations of high spatio-temporal resolution to enable operational landslide monitoring. In this research, the construction of digital elevation models (DEMs) and orthomosaics from UAV imagery is achieved using structure-from-motion (SfM) photogrammetric procedures. The study examines the additional value that the morphological attribute of openness, amongst others, can provide to surface deformation analysis. Image-cross-correlation functions and DEM subtraction techniques are applied to the SfM outputs. Through the proposed integrated analysis, the automated quantification of a landslide's motion over time is demonstrated, with implications for the wider interpretation of landslide kinematics via UAV surveys.

Traffic flow analysis is fundamental for urban planning and management of road traffic infrastructure. Automatic number plate recognition (ANPR) systems are conventional methods for vehicle detection and travel times estimation. However, such systems are specifically focused on car plates, providing a limited extent of road users. The advance of open-source deep learning convolutional neural networks (CNN) in combination with freely-available closed-circuit television (CCTV) datasets have offered the opportunities for detection and classification of various road users. The research, presented here, aims to analyse traffic flow patterns through fine-tuning pre-trained CNN models on domain-specific low quality imagery, as captured in various weather conditions and seasons of the year 2018. Such imagery is collected from the North East Combined Authority (NECA) Travel and Transport Data, Newcastle upon Tyne, UK. Results show that the fine-tuned MobileNet model with 98.2 % precision, 58.5 % recall and 73.4 % harmonic mean could potentially be used for a real time traffic monitoring application with big data, due to its fast performance. Compared to MobileNet, the fine-tuned Faster region proposal R-CNN model, providing a better harmonic mean (80.4 %), recall (68.8 %) and more accurate estimations of car units, could be used for traffic analysis applications that demand higher accuracy than speed. This research ultimately exploits machine learning alogrithms for a wider understanding of traffic congestion and disruption under social events and extreme weather conditions.

Moisture-induced landslides are a global geohazard; mitigating the risk posed by landslides requires an understanding of the hydrological and geological conditions present within a given slope. Recently, numerous geophysical studies have been attempted to characterise slow-moving landslides, with an emphasis on developing geoelectrical methods as a hydrological monitoring tool. However, landslides pose specific challenges for processing geoelectrical data in long-term monitoring contexts as the sensor arrays can move with slope movements. Here we present an approach for processing long-term (over 8 years) geoelectrical monitoring data from an active slow-moving landslide, Hollin Hill, situated in Lias rocks in the southern Howardian Hills, UK. These slope movements distorted the initial setup of the monitoring array and need to be incorporated into a time-lapse resistivity processing workflow to avoid imaging artefacts. We retrospectively sourced seven digital terrain models to inform the topography of our imaging volumes, which were acquired by either Unmanned Aerial Vehicle (UAV)-based photogrammetry or terrestrial laser ranging systems. An irregular grid of wooden pegs was periodically surveyed with a global position system, from which distortions to the terrain model and electrode positions can be modelled with thin plate splines. In order to effectively model the time-series electrical resistivity images, a baseline constraint is applied within the inversion scheme; the result of the study is a time-lapse series of resistivity volumes which also incorporate slope movements. The workflow presented here should be adaptable for other studies focussed on geophysical/geotechnical monitoring of unstable slopes.

Landslides are hazardous events with often disastrous consequences. Monitoring landslides with observations of high spatio-temporal resolution can help mitigate such hazards. Mini unmanned aerial vehicles (UAVs) complemented by structure-from-motion (SfM) photogrammetry and modern per-pixel image matching algorithms can deliver a time-series of landslide elevation models in an automated and inexpensive way. This research investigates the potential of a mini UAV, equipped with a Panasonic Lumix DMC-LX5 compact camera, to provide surface deformations at acceptable levels of accuracy for landslide assessment. The study adopts a self-calibrating bundle adjustment-SfM pipeline using ground control points (GCPs). It evaluates misalignment biases and unresolved systematic errors that are transferred through the SfM process into the derived elevation models. To cross-validate the research outputs, results are compared to benchmark observations obtained by standard surveying techniques. The data is collected with 6 cm ground sample distance (GSD) and is shown to achieve planimetric and vertical accuracy of a few centimetres at independent check points (ICPs). The co-registration error of the generated elevation models is also examined in areas of stable terrain. Through this error assessment, the study estimates that the vertical sensitivity to real terrain change of the tested landslide is equal to 9 cm.