The aim of this paper is to classify and segment roofs using vertical aerial imagery to generate three-dimensional (3D) models. Such models can be used, for example, to evaluate the rainfall runoff from properties for rainwater harvesting and in assessing solar energy and roof insulation options. Aerial orthophotos and building footprints are used to extract individual roofs and bounding boxes, which are then fed into one neural network for classification and then another for segmentation. The approach initially implements transfer learning on a pre-trained VGG16 model. The first step achieves an accuracy of 95.39% as well as a F1 score of 95%. The classified images are segmented using a fully convolutional network semantic segmentation model. The mask of the segmented roof planes is used to extract the coordinates of the roof edges and the nexus points using the Harris corner detector algorithm. The coordinates of the corners are then used to plot a 3D Level of Detail 2 (LOD2) representation of the building and the roof height is determined by calculating the maximum and minimum height of a Digital Surface Model LiDAR point cloud and known building height data. Subsequently the wireframe plot can be used to compute the roof area. This model achieved an accuracy of 80.2%, 96.1%, 96.0%, 85.1% and 91.1% for flat, hip, gable, cross-hip and mansard roofs, respectively.

<p><span>In order to enable community groups and other interested parties to evaluate the effects of flood management, water conservation and other hydrological issues, better localised mapping is required.  Although some maps are publicly available many are behind paywalls, especially those with three dimensional features.  In this study London is used as a test case to evaluate, machine learning and rules-based approaches with opensource maps and LiDAR data to create more accurate representations (LOD2) of small-scale areas.  Machine learning is particularly well suited to the recognition of local repetitive features like building roofs and trees, while roads can be identified and mapped best using a faster rules-based approach. </span></p><p><span>In order to create a useful LOD2 representation, a user interface, processing rules manipulation and assumption editor have all been incorporated. Features like randomly assigning sub terrain features (basements) - using Monte-Carlo methods - and artificial sewage representation enable the user to grow these models from opensource data into useful model inputs. This project is aimed at local scale hydrological modelling, rainfall runoff analysis and other local planning applications. </span></p><p><span> </span></p><p><span>The goal is to provide turn-key data processing for small scale modelling, which should help advance the installation of SuDs and other water management solutions, as well as having broader uses. The method is designed to enable fast and accurate representations of small-scale features (1 hectare to 1km<sup>2</sup>), with larger scale applications planned for future work.  This work forms part of the CAMELLIA project (Community Water Management for a Liveable London) and aims to provide useful tools for local scale modeller and possibly the larger scale industry/scientific user. </span></p>

<div> <p class="paragraph"><span class="normaltextrun"><span lang="EN-US">Climate change associated sea level increases and projected growth in global water consumption of about 1% per year (WWAP, 2018) are expected to place further demands on already heavily utilized coastal groundwater supplies. Water stress is anticipated to become more critical over the next decades (Werner and Simmons, 2009). Society’s over-reliance on coastal freshwater abstraction had led to an increased threat of Saline Intrusion (SI). In spite of these challenges, no widely applicable methods of tracking saline fronts in the subsurface exist, even though this capability could prove critical to stopping over abstraction (pumping) before SI occurs; o</span></span><span class="normaltextrun">bservational boreholes offer a limited warning, and resistivity imaging is often too expensive and logistically infeasible,</span><span class="normaltextrun"><span lang="EN-US"> (MacAllister et al. 2016). An alternative approach to detecting imminent SI is needed. The ongoing goal of this work is to develop a robust and low-cost method of tracking SI in the sub-surfaces.</span></span><span class="eop"> </span></p> </div> <div> <p class="paragraph"><span class="normaltextrun"><span lang="EN-US">Naturally occurring voltages, known as Self Potential (SP), occur when pressure and concentration gradients in the subsurface cause ion separations </span></span><span class="normaltextrun">(Jackson et al., 2012)</span><span class="normaltextrun"><span lang="EN-US">. </span></span><span class="normaltextrun"> SP can be used to track SI, so long as the signal source mechanism is understood.</span> <span class="normaltextrun">There are two key sources of SP widely encountered in hydrology, those induced by pressure, electro-kinetic potentials (V</span><span class="normaltextrun"><sub>EK</sub></span><span class="normaltextrun">), and exclusion-diffusion potentials (V</span><span class="normaltextrun"><sub>ED</sub></span><span class="normaltextrun">), due to ion concentration gradients moving through the subsurface. </span><span class="eop"> </span></p> </div> <div> <p class="paragraph"><span class="normaltextrun"><span lang="EN-US">SP signals are generated relative to static reference electrodes, offering a signal reading per electrode. However, these signals drift over time making interpretation and comparison challenging. We present findings and insights of an investigation using travelling SP electrodes, moving vertically inside boreholes or wells, to generate SP profiles. Results offer new insights into relationships between SP and SI when logged over time. Profiles taken over the last year at a variety of coastal and inland sites in the UK build upon results from a controlled pumping experiment in Northern Ireland, completed in 2020 and which attempted to interpret these patterns and signals through machine learning. Filtering out background noise sources, (such as electrical interferance, tides, Magneto Telluric effects etc.) has allowed signatures to be more confidently generated and related SI under contrasting hydrogeological regimes. This novel methodology and initial findings are presented and the scope for widely application of the method discussed.</span></span></p> <p class="paragraph"> </p> <p class="paragraph"><strong><span class="normaltextrun">References</span></strong></p> <p>Jackson, M. D., et al.   (2012). Measurements of spontaneous potential in chalk with application to aquifer characterisation in the southern UK quarterly. <em>J. Eng. Geol. Hydrogeol</em>.</p> <p>MacAllister, et al. (2016), Tidal influence on self-potential measurements, <em>Journal Geophysical Research Solid Earth.</em></p> <p>Werner, A.D., Simmons, C.T., (2009), Impact of sea‐level rise on seawater intrusion in coastal aquifers. Ground Water.</p> <p>WWAP, (2018), The United Nations World Water Development Report 2018: Nature-Based Solutions for Water. Paris, UNESCO.</p> <p class="paragraph"> </p> </div>

<p>Due to greater ground-water abstraction, rising demand for water, possible reductions in recharge rates and rising sea levels, costal aquifers are under ever increasing threat of Saline Intrusion (SI) (Mehdizadeh, 2019). Though the mechanismins of SI have long been understood, the ability to monitor and warn in advance of the ingress of saline water into costal aquifers has remained costly and complex (Graham, 2018). The work reported here describes initial efforts to develop and results from, a vertically profiling Self Potential (SP) device. The device was used to monitoring the position of a well parametrized saline front in a costal aquifer, located on Benone Strand, Co. Derry,  on the northern tip of Northern Ireland, UK, as part of the SALine INtrusion in coastal Aquifers project.</p><p>Naturally arising voltages, Self Potential (SP), are formed when pressure and concentration gradients move though the subsurface. The gradients cause ion separations, which create electrical potentials and a flow of electrons in order to maintain electrical neutrality. The SP signals (usually in the millivolt range) can be detected, relatively inexpensively (in comparison to resistivity imagining) with reference electrodes and a high impedance voltage logger. The positioning of the electrodes is key as it has only been possible, until now, to measure the voltage between two points. There are two key types of SP, in hydrology, electro-kinetic potentials (V<sub>EK</sub>), due to differential flow velocities, and exclusion-diffusion potentials (V<sub>ED</sub>), due to ion concentration gradients with different mobilities. Understanding the source mechanims in these voltages is complex, but evolving. Previous work has shown that self-potential rises before a saline breakthrough into a borehole (Graham, 2018).</p><p>A novel vertically travelling (or trolling) SP electrode was repeatedly used in a number of satellite boreholes during a pumping test; in order to look at the changes in the vertical gradient of SP. The pumping test took place over three days, during which initially fresh water was abstracted from the main pumping well. Resistivity imagine was used as a benchmark. It was shown that the vertical SP profile changed as the salt content of the pumped water increased (i.e. the saline front moved inland). This change in SP could not be explained by pressure changes – gradients of 50mV inside a single borehole were observed. The data showed SP profiles that varied widely before, during and after the pumping test, as saline water is drawn progressively towards the pumping well, offering far more data than a single stationary electrode. Demonstrating that these signals change in advance of the saltwater arriving at the pumping well, but also that this method could be used as an inexpensive way to safeguard costal aquifers in the future.</p><p><strong>References</strong></p><p>Graham, M. T. (2018). Self-Potential as a Predictor of Seawater Intrusion in Coastal. <em>Water Resources Research</em>.</p><p>MacAllister, D. a. (2016). Tidal influence on self-potential measurements. <em>Journal of Geophysical Research: Solid Earth</em>.</p><p>Mehdizadeh, S. a. (2019). Abstraction, desalination and recharge method to control seawater intrusion into unconfined coastal aquifers. <em>Global Journal of Environmental Science and Management, 5</em>, 107-118.</p><p> </p><p> </p>

No abstract