OSGeo:UK

OSGeo UK Local Chapter

Sponsors

Venue & Party

Geovation Astun

Gold

Esri UK

Silver

Carto GeoCat Geolytix WRLD

Bronze

Axis Maps
Google
Scisys

FOSS4G UK 2018 : Abstracts

Thursday

Title Main Presenter Room
Welcome James Milner, 3D Repo & Laura Alderson, Geovation Auditorium
OSGeo: A Brief Introduction Jo Cook, OSGeo:UK Auditorium
Keynote: Spatial Data Infrastructures: in the age of Docker and Microservices Joana Simoes, Geocat Auditorium
QGIS: A Sustainable Open Source Saber Razmjooei, Lutra Consulting Auditorium
Publishing MapAction Maps: A QGIS Plugin Ant Scott, MapAction Auditorium
QGIS Anywhere Martin Dobias, Lutra Consulting Auditorium
Supercharging QGIS Workflow with PostGIS Andrew Bell, Geoacent Auditorium
Spatial Web Services Serbulent Ocal, BT/EE Auditorium
XYZ - Modular Web Mapping, Reporting and Development for Consultants Dennis Bauszus, Geolytix Auditorium
Load Balancing & Load Testing Strategies for Web Map Servers Ian Turton, Astun Technology Auditorium
Vector Tile Rendering with Open Source Tools and Open Data Dave Barter, Nautoguide Ltd Auditorium
Keynote: Extracting intelligent information from aerial images using machine learning Mathilde Ørstavik, Norkart Auditorium
Close: Owt for Nowt and Tuppence Change Tom Chadwin, OSGeo/Northumberland National Park Auditorium

Friday

Title Main Presenter Room
CARTO as a Platform Jorge Sanz, Carto Auditorium
Open Drones Thomas Starnes, RSPB Auditorium
Pirate Maps: Portable Maps on the Raspberry Pi Ian Turton, Astun Technology Workspace
GIS: Software or Solution John Byrne, Mapail Workspace
3D Web GIS with the I3S Standard Dan Cronin & Richard Mumford, ESRI UK Hub Space
Multispectral Augmented Classification of 3D Dense Point Cloud Using Machine Learning Sheikh Fakhar Khalid, SenSat Hub Space
Migrating Accessibility Measures to Open Source Richard Williams, University of South Wales Boardroom
Building GeoViewer 2: OS Internal Data Viewer Tim Martin Boardroom
Enabling Transformation Through Efficient Use of Earth Observation Data Steve Wilkinson & Pascal Coulon, Defra/SCISYS Auditorium
Satellite Data, Raw to Useful Matt Debont & Lynn Heeley, JNCC/DEFRA Auditotium
Pronto Raster: A New C++ Library for Map Algebra Alex Hagen-Zanker, University of Surrey Auditorium
Using non-PostGIS Data with Foreign Data Wrapper Aileen Heal, Astun Technology Workspace
Show Off Your Spatial! You Hub Space
The Importance of Design in Geo: GeoDataViz Toolkit Charley Glynn, Ordnance Survey Auditorium
Better Digital Cartography with OpenLayers 4 Oliver O’Brian, UCL Auditorium
Lightsaber Maps: Using Draw Effects and Blend Modes in QGIS Tom Armitage, EDINA, University of Edinburgh Auditorium
Visualising School Catchment Areas Ross McDonald, Angus Council Auditorium
Can Free and Open Source Geospatial Solutions Replace Proprietary Software in a Business Environment? Chris Gale, ONS Hub Space
Is Europe Losing Out on Open Source Big Geospatial Data Analytics? Marc Vloemans, LocationTech/Eclipse Foundation Hub Space
Keynote: Geospatial Industry Trends Peter Batty, Ubisense Auditorium
Close James Milner, 3D Repo Auditorium

Welcome

James Milner, 3D Repo & Laura Alderson, Geovation

Welcome


OSGeo: A Brief Introduction

Jo Cook, OSGeo:UK

An introduction to OSGeo and what it does.


Keynote: Spatial Data Infrastructures: in the age of Docker and Microservices

Joana Simoes, Geocat

SDIs are a relatively modern “invention”, which leverages the value of spatial information, but it does not come without challenges. One of these challenges is that they can be really hard to setup.

In this talk, I will address this challenge with one question “How can we ease the provision of SDIs?” In my quest to answer this question, I will take you on a journey through recent technologies such as OS virtualization, cloud and distributed computing.


QGIS: A Sustainable Open Source

Saber Razmjooei, Lutra Consulting

QGIS has come a long way in the past 15 years. It has now become one of the main GIS desktop application alongside proprietary applications.

To understand how this has been achieved, the presentation will offer an insight to the success and great community behind the project. It will also highlight the future challenges and how those can be addressed.


Publishing MapAction Maps: A QGIS Plugin

Ant Scott, MapAction

In the course of its humanitarian response work, MapAction produces a stream of maps which are published to its website and to other sites. These are available as PDF and JPG, and are accompanied by an XML metadata file which is used by the CKAN-based map catalogue. MapAction currently uses its own add-in to generate this fileset from ArcGIS, which is its primary desktop production tool. To support map production in QGIS as an alternative to Arc, MapAction have developed a QGIS plugin which mirrors the current workflow, supporting the use of QGIS in deployments when needed.

The talk will cover the process of building the plugin, and the QGIS templates it uses, describe some of the issues encountered along the way, and demonstrate the plugin in use.


QGIS Anywhere

Martin Dobias, Lutra Consulting

QGIS is not only a desktop application. The application has evolved to serve maps and also on mobile/table devices as survey application.

In this presentation, we will present the full stack of QGIS applications from survey in the field to publishing data on the web. It will also detail our works on QGIS mobile https://github.com/qgis/QGIS-Enhancement-Proposals/issues/109 which introduces a new framework for developing mobile applications using QGIS libraries.


Supercharging QGIS Workflow with PostGIS

Andrew Bell, Geoacent

The presentation provides an overview of how PostgreSQL/ PostGIS can be used to semi-automate and enhance business workflow through QGIS.

In use of working examples the session will look at the use database functions to build relationships between datasets and semi automate processes, this combined with QGIS form architecture as a configurable interface. The use of PostgreSQL row level and column based security will demonstrate how data distribution to QGIS users in corporate environment can be refined to meet specific access requirements.


Spatial Web Services

Serbulent Ocal, BT/EE

Telecommunications industry is adapting in a major way as new trends have emerged in software and services such as 5G, Artificial Intelligence (AI), Cloud, Internet of Things (IoT), Smart Cities, VR/AR and Open Source.

The huge expansion and development of telecommunications industry means that the landscape has become much more complex than it was ten or twenty years ago. There are new ways of working as stakeholders need to cooperate and share information more openly. So that there is a need of web services to provide smart networked communications environment for co-operation and sharing information.

In order to fully understand what a web service is and to appreciate what it can do for mobile technologies, with an emphasis on future 5G (fifth generation); there is a need to understand the software development challenges in mobile technology such as Java EE, open source, APIs (open – private), REST APIs & Spatial Web Servers (Geoserver&Geowebcahe) and client components.

This presentation will first outline these challenges with demonstrating some examples. Then, based on E-Business (Telecommunications), it will provide an overview of how we created spatial web services (Geocoding & Coverage) as REST Style Web Services.

As a working example, I’d like to introduce EE Coverage Checker Application, and share my experiences while creating these REST Style Web Services.


XYZ - Modular Web Mapping, Reporting and Development for Consultants

Dennis Bauszus, Geolytix

Being a developer for a location intelligence consultant it became quickly apparent that going open source is the only feasible approach in building a platform to support our data products and consultancy services.

XYZ is an open source (MIT licence) modular node framework which may serve secure browser applications and RESTful web services.

The framework aims to be a lightweight middleware and browser interface to allow secure access to a range of data sources without the need for additional services. Dedicated UX being mostly an industry standard in 2018 the focus is now on developer experience (DX). Addressing thus the needs of developers who prefer a modular framework over an application builder.

Some unique features will be presented along with an overview of the features which are currently designed.


Load Balancing & Load Testing Strategies for Web Map Servers

Ian Turton, Astun Technology

When deploying a Web Map Server (WMS) or Tile Cache (WMTS/TMS) system for your organisation the question you most want to know is how many users can it support at once. It is best to be able to answer this question before the critical incident that it will be used to support is upon you.

This talk will first outline some of the possible configurations of WMS and WMTS servers and caches that can be used to meet possible peaks in demand for map tiles.

The talk will then move onto the ways you can easily test various configurations to give you a feel for how many users you can support in day to day and emergency usage.


Vector Tile Rendering with Open Source Tools and Open Data

Dave Barter, Nautoguide Ltd

I have been on a journey to create a fantastic client side map rendering experience using vector tiles and opensource mapping toolkits. During this talk I will explain the advantages and pitfalls of using vector tiles and give a real word example of rendering client side maps using Ordnance Survey open data. The talk will include the steps required to create and host tiles along with all source code and a set of tools we’ve created to do this and are realising as open source.


Keynote: Extracting intelligent information from aerial images using machine learning

Mathilde Ørstavik, Norkart

The revolution within machine learning has given rise to new, real-life, applications, especially within computer vision. This talk will focus on not just how these new machine learning techniques work, but how they can be exploited for geospatial purposes. When analyzing georeferenced images, you will not only learn what the images are depicting but you can also derive where in the world the recognized objects exist. This gives a whole new meaning to the extracted data.

At Norkart we have combined our knowledge of geographical data with the newest machine learning techniques. We have tested to what extent we can extract geographical information from aerial and satellite imagery by using deep learning. Several convolutional neural networks for segmentation have been implemented and tested on multispectral geospatial images as well as traditional RGB images to improve the accuracy further.

Is it possible to extract sufficiently accurate building data in order to automatically calculate the effect of solar panels? Can you discover buildings that are built illegally? Can you extract information to helps risk analysis for insurance purposes? These are questions that will be discussed further, based on our results.


Close: Owt for Nowt and Tuppence Change

Tom Chadwin, OSGeo/Northumberland National Park

Blogpost


CARTO as a Platform

Jorge Sanz, Carto

CARTO is a successful service and Open Source product stack used by organizations of all sizes and verticals. Its portfolio is mainly divided into two main products: BUILDER and ENGINE. This talk is about the latter, exploring how to use the CARTO platform to develop your geospatial applications with a complete set of services, APIs, and client components.

First, we’ll focus on the APIs and Location Data Services offered by CARTO and our plans to evolve them in 2018. Secondly, we’ll cover our JavaScript SDK: CARTO.js. This brand new library offers a different approach; more developer focused and low-level. Finally, we will discuss CARTO for Python developers, with our Python SDK and the CARTOFrames library we’ve created on top to enable direct integration with Pandas and Jupyter Notebook, the increasingly popular environment for Data Scientists.

Slides


Open Drones

Thomas Starnes, RSPB

Here we provide a high level workflow in which we use OpenDroneMap to generate high resolution orthographic aerial imagery using commercial standard grade drones. We then upload this imagery to OpenAerialMap - an online platform hosting openly licensed imagery and map services - and subsequently make edits to OpenStreetMap. Finally we’ll provide some eye candy; using derivative drone products to render 3D elevation models in a web browser using qgis2threejs.


Pirate Maps: Portable Maps on the Raspberry Pi

Ian Turton, Astun Technology

This talk will report on some initial experiments on using a Raspberry Pi to create local caches of maps to speed up GIS systems with slow network connections or low bandwidth.

The Raspberry Pi is a series of small single-board computers developed in the United Kingdom by the Raspberry Pi Foundation to promote the teaching of basic computer science in schools and in developing countries. The original model became far more popular than anticipated, selling outside of its target market for uses such as robotics. Peripherals (including keyboards, mice and cases) are not included with the Raspberry Pi. The low cost and simplicity of the Raspberry Pi makes it ideal to use in situations when a simple plug and play machine is required.


GIS: Software or Solution

John Byrne, Mapail

GIS is now a software engineering process as much as any other discipline. As such how it is handled can and should be done in a consistent and repeatable process. This talk will discuss how we have taken a GIS solution containing both open source and commercial code and migrate it through a repeatable and testable environment build process from development to live.


3D Web GIS with the I3S Standard

Dan Cronin & Richard Mumford, ESRI UK

The Indexed 3D Scene (I3S) format can store and serve a range of massively large 3D GIS datasets over the web. The OGC adopted I3S as a Community Standard in 2017, providing a new open specification for sharing 3D content. This session will showcase data, examine the underlying technology and consider the place of I3S in the broader 3D GIS community.


Multispectral Augmented Classification of 3D Dense Point Cloud Using Machine Learning

Sheikh Fakhar Khalid, SenSat

3D points clouds are traditionally generated using two common techniques; LiDAR and SfM (Structure for Motion) based photogrammetry. Full waveform LiDAR tends to use single wavelength, restricting their use to topographical analysis only. Even though LiDAR data can be augmented with RGB values from a visible range camera, it is not an ideal case for unravelling spectral classes in order to autonomously detect objects on the terrain. Recent years have seen emergence of multispectral LiDAR platforms which allows spectral variations to be analysed in addition to topographical information, but they tend to require large UAV platforms.

This exploratory paper proposes an augmentation machine learning approach to fuse multispectral satellite data and supplementary open source spatial data with SfM based point cloud to classify dense urban regions. An object-based image analysis classification technique was used to traditionally classify multispectral and multiscale Sentinel-2 MSI imagery for Cambridge. The 10m and 20m spatial resolution classified images were then augmented and enhanced using 2.5cm GSD (Ground Sampling Distance) aerial orthorectified imagery. The resulting information was then fused with dense point cloud data generated using SfM. This allowed the point cloud in 3D space to be enhanced with ancillary sub-resolution multispectral information. A nearest-neighbour algorithm was then used to reclassify objects in 3D space using the augmented point clouds. The early exploratory results are promising; however, the object classification accuracy can be improved in both type I and type II errors.

A ground validation schema was also adopted to measure accuracies of the results and open source data such as existing building footprint were used to tighten up the classification outputs.


Migrating Accessibility Measures to Open Source

Richard Williams, University of South Wales

There is a large literature that has explored the use of ‘floating catchment area’ (FCA) techniques in measuring access to a range of primary and secondary health services. This paper describes the development of open source tools which incorporate FCA techniques initially applied in the context of mapping and modelling access to mobile cancer services. Tests completed on the open source tool allow it be used to compare to an existing proprietary tool developed for use in ArcGIS. This study includes reflections on the underlying data used within these models as well as the software itself in order to better understand differences in outputs and computational speed when performing FCA analyses in an open source environment.


Building GeoViewer 2: OS Internal Data Viewer

Tim Martin, Ordnance Survey

This talk will show off GeoViewer2, OS’s new internal data viewer, which is used by teams across OS to view, query and access the latest and archive versions of OS datasets. During the talk we will discuss the architecture and the FOSS4G we use for loading (NodeJS & GDAL/OGR2OGR), storing (PostgreSQL/PostGIS) and viewing the data (QGIS).


Enabling Transformation Through, Efficient Use of Earth Observation Data

Steve Wilkinson & Pascal Coulon, Defra/SCISYS

Defra wanted to build on two earlier Proofs of Concept to deliver an infrastructure to provide access to analysis ready EO data and to demonstrate how this can be applied to support operational areas and to provide data feeds across Defra and related partners. Defra Partners include for example the Defra Earth Observation Centre of Excellence (EOCoE), a collaborative group of Defra bodies and external organisations.

With the support from their delivery partner SCISYS, a cloud based open architecture platform using solely open source components has been designed. Based mainly OSGeo components, the project is delivering a scalable and extensible platform for the delivery of Sentinel 1 and Sentinel 2 OGC feeds.

This presentation will address some of the key step employed to automate the retrieval and processing of raw sentinel 1 and sentinel 2. In the second part this presentation will demonstrate how a clustered environment coupled with elastic computing enabled the delivery of a scalable platform.


Satellite Data, Raw to Useful

Matt Debont & Lynn Heeley, JNCC/DEFRA

There are now 6 Sentinel Satellites in orbit beaming down TB’s of free data, but what can we actually do with it? Following on from a potential talk ““Enabling Transformation through efficient use of Earth Observation data”” around what we at JNCC / DEFRA are trying to do with Sentinel data. From collating the data to processing it into a form that is more immediately useful using AWS cloud services and open source tools to how we envision using that data to improve governance on the environment and further use by academia and the general public.

The talk can be roughly split into two parts, the first part mostly around processing the data using docker containers and AWS Batch to contain the logic, orchestrated by Luigi (a workflow management system).

The second part concerning what we are actually going to be doing with this data, and how it can be used by others. I would like to talk about some of the data products that have already been developed as well as some more embryonic ones and how they are created. This could also take on a more free form discussion about possible ways the data could be used once processed to a standard form.


Pronto Raster: A New C++ Library for Map Algebra

Alex Hagen-Zanker, University of Surrey

The Pronto Raster library is a C++ library for Map Algebra operations. Map Algebra is a long established conceptual framework for geographical data analysis [1]. It is a versatile and highly generic framework, classifying local, focal and zonal operations. Many tools and libraries exist that offer Map Algebra functionality. However existing libraries are limited by offering specific tools with little room for customization; Or, when they are customizable, it is at a substantial loss of efficiency. The Pronto Raster library aims to overcome this and provides an efficient computational framework that allows local, focal and zonal operations to be applied with user-specified functions using standard and straightforward C++ idioms.

In particular it achieves this efficiency by combining the Expression Template technique and the Range concept. By exposing raster data as ranges, values in a raster can be iterated over using the C++11 range-based for-loop. It hides the complexity of raster data access, such as caching blocks of data, behind the most standard and basic interface. The Expression Template technique means that operations on raster data can be performed without unnecessarily creating temporary datasets for intermediate results, making operations more efficient and robust than existing approaches. A number of elementary spatial operators provide transformed views of existing raster datasets, that also do not require creating temporary datasets, such as iterating over a subraster, or iterating over cells in a raster at a spatial offset. These elementary spatial transforms can be combined into more complex spatial operations - all without creating unnecessary temporary datasets - including highly efficient moving window analysis.

The library is intended to especially useful for environmental modelling, including Map Algebra operations, Cellular Automata modelling, spatial statistics and multi-scale analysis.

[1] Tomlin CD. Geographic information systems and cartographic modeling. Prentice Hall; 1990.


Using non-PostGIS Data with Foreign Data Wrapper

Aileen Heal, Astun Technology

PostgreSQL foreign data wrappers https://wiki.postgresql.org/wiki/Foreign_data_wrappers are a way of accessing (both reading and writing) data held outside your PostgreSQL database and making them appear to be held within your PostgreSQL database.

There are a number of foreign data wrappers (FDW) extensions which support spatial data.

This presentation will show how you can install some of these extensions and show how to read/write to spatial data in different databases (including a different PostGIS database and an Oracle database) and files e.g. ESRI shape, ESRI file geodatabases, MapInfo TAB files etc. or even web services e.g. WFS.

Slides


Show Off Your Spatial!

You

Suggest a lightning talk on Thursday.


The Importance of Design in Geo: GeoDataViz Toolkit

Charley Glynn, Ordnance Survey

There was an apparent lack of content around visualisation at the last FOSS4G UK event, so I want to fly the flag for the importance of design within geo. 2017 saw a continuation in the boom of data visualisation and cartography with data journalists often leading the way with beautiful and powerful graphic-led storytelling. Open source software underpins a lot of this great work and open data is enabling greater insights and truths. This talk will explore why design is more important than ever for geo and offer some advice and best practices.

There are many, many open datasets, and an increasingly great selection of FOSS4G tools available these days. There are also some great learning materials available on the web and a real desire to share knowledge for the greater good. Last year the GeoDataViz team at Ordnance Survey created a toolkit of assets and resources to help people communicate their data effectively through the design of compelling visuals. Design assets including colours and symbols are supported by best practice, tips and workshop material and it is all freely available for anybody to use. I’d like to introduce the GeoDataViz Toolkit and offer some tips for improving your maps and data visualisations.


Better Digital Cartography with OpenLayers 4

Oliver O’Brian, UCL

This talk showcases, with simple code examples, uses of new digital cartography techniques in recent visualisations of data at the Consumer Data Research Centre, focusing particularly on some of the new features in the OpenLayers 4 Javascript mapping framework. Examples include live keys, spatial analysis (with JSTS), automated decluttering, triangular colour ramps, styling labels, map downloads, UTF grids and streamlined data input with GeoJSON files.


Lightsaber Maps: Using Draw Effects and Blend Modes in QGIS

Tom Armitage, EDINA, University of Edinburgh

QGIS has gained increasingly advanced features for rendering spatial data. This talk highlights how it is possible to use these features to create very high quality cartographic outputs. Draw effects such as inner and outerglows as well as blurs can be utilised in a range of ways to create visually arresting graphics. Blend modes are also available, allowing multiple layers of data to be combined without the usual dulling effect of transparencies.

This talk will go through some of the myriad ways these effects can be used and combined to create maps with impact. There will be some examples of the good, bad and ugly results that can be acheived.


Visualising School Catchment Areas

Ross McDonald, Angus Council

A case study on how spatial data was used to help visualise school catchment areas in different ways. From simple maps made in QGIS to interactive 3D maps in a browser and 3D scenes made in Blender.

Slides


Can Free and Open Source Geospatial Solutions Replace Proprietary Software in a Business Environment?

Chris Gale, ONS

Can the Office for National Statistics (ONS) implement free and open source geospatial solutions to completely replace proprietary GIS software? There are a number of challenges ONS would face if it were to ever transition fully away from proprietary GIS solutions. A non-exhaustive list of these challenges includes the following:

We recognise that each of these challenges can be met individually by free and open source geospatial solutions. However, we wish to engage with users of free and open source geospatial solutions to understand what options ONS has if it wants to address these challenges in the most harmonised way possible.


Is Europe Losing Out on Open Source Big Geospatial Data Analytics?

Marc Vloemans, LocationTech/Eclipse Foundation

Big Spatial Data technology based on geospatial attributes for the Cloud is relatively new to the developer community and organisational ecosystem. To find the latest proven software one has to look across the Atlantic Ocean, where a suite of specialised open spatial solutions is emerging.


Peter Batty, Ubisense

This talk will discuss current and future trends in the geospatial industry. Peter will talk about the increasing importance of open source and share some of his experiences of building large scale enterprise geospatial applications on open source foundations over the past ten years. He will look at upcoming changes in the broader geospatial industry which will be dramatic, especially in terms of how we capture and maintain data about the world around us.


Close

James Milner, 3D Repo

Close.