Three-Dimensional Analysis of Forest Structure and Terrain Using LIDAR Technology
(Phase I – RES#50)
Forests represent an important economic resource throughout Canada, accounting for $39.7 billion in export trade in 1998 alone (NRCan, 1999). Forests also account for significant exchanges of energy, sensible heat, water, CO2 and trace gases with the lower atmosphere. For economic and environmental reasons, it is critical to measure and understand the spatial organization of forest ecosystems. (Project Final Report, 2002)
In Phase I, a team of researchers used a new laser-based technology (LiDAR) to provide more accurate measurements at the level of the forest canopy. The project focused on the evaluation of algorithms to estimate structural and biophysical variables of forests and terrain characteristics. “A high resolution LIDAR-derived Digital Elevation Model (DEM) of the TLW was acquired that out-performed other data sources. The LiDAR DEM was used with air photos to produce three-dimensional information on the forest canopy.” (Environment Canada, 2002)
One key result of the Phase I project was recruiting Dr. Kevin Lim, who carried through this line of research. He founded a small consulting startup (LimGeomatics) working on LiDAR applications to forestry. Over a ten-year-period, LiDAR has matured as a key technology in geomatics and, by Phase IV, some of the same researchers validated the application of the original concepts on large forest holdings.
Generalization and Multiple Representations for On-demand Map Production and Delivery
(Phase II – MNG#BED)
This project led to new concepts, methods and tools combing cartographic generalization and multiple representations at different levels. Among other things, the team developed:
- A data warehouse that integrates several data sources into a multiple representation structure (MR). Building this data warehouse required an ontological analysis of the different concepts of each data source (to match semantically equivalent concepts) followed by a geometric data-matching process to link different representations of a same object;
- New concepts of geometric patterns to reduce the data volume of multiple representation databases (MRDB) and facilitate automatic generalization processes;
- New concepts of Self-Generalizing Objects (SGO) that combine geometric patterns, generalization algorithms and spatial integrity constraints to create a complete generalization behaviour that can be used by any occurrence of a database in a scale-reducing process;
- A new preliminary multi-scale data acquisition method based on predefined concepts allowing the creation of MRDB from the acquisition phase. This method bypasses the labour-intensive and complex integration process;
- A new method for evaluating the quality of generalized maps based on a unique framework that allows the calculation of a single metric that quantifies the quality of the map;
- A new method for evaluating the impact of several representations of the data at different scales on reasoning processes
- A preliminary framework for data capture in a MRDB context. (Project Final report, 2005)
Under Dr. Yvan Bédard’s leadership, the project team developed two applications to test and validate:
- UMapIT, an on-demand web-mapping application based on a specific MR datamart. This application is based on efficient use of the stored multiplicities (geometric, semantic and graphic). It offers the user a level of flexibility not provided until now by web-mapping applications;
- SIGERT, a geo-located web services for recreo-tourism domains based on another MR datamart. This application uses an innovative multi-agent system to overcome crowding problems that may arise when displaying a map.
On-demand mapping is defined as a customer-driven process that creates tailor-made maps, often in a unique ready-to-print copy. (Project Final report, 2005)
Integrated Modelling of Juvenile Atlantic Salmon Movement and Physical Habitat in Fluvial and Estuarine Environments
(Phase III – ISLMDFM14)
The economic value of wild Atlantic salmon stems largely from the sport fishery, worth tens of millions of dollars annually. The species is, however, in decline across its natural range, prompting a call to action to resource managers and the science community. In Phase III, a GEOIDE team (SLMDFM-14, led by Dr. Julian Dodson) adopted an integrated approach to salmon habitat from headwaters to estuaries; mobilizing fluvial geomorphology, biology, and geomatics technology. One component deployed arrays of geographically positioned acoustic hydrophones to continuously monitor the behaviour of acoustically tagged fish. (These are active tags that emit a signal based on a limited-life battery.) This technology permitted the team to identify the factors influencing the migration and swimming depth of kelt at a landscape scale within the York Estuary and Gaspé Bay (Québec, Canada). Additional receivers located in the Strait of Belle Isle (Newfoundland, Canada) allowed them to document for the first time kelt leaving the Gulf of St. Lawrence.
At the local scale, two general migration patterns were observed. Approximately one-third of the kelt showed a highly oriented migration across the estuary-coastal embayment with little diving, rapidly moving out of Gaspé Bay. The remaining kelt showed repeated changes in direction and multiple dives before moving out of Gaspé Bay. There was long-term residence (typically several weeks) in the river delta suggesting active feeding to regain condition following the winter fast. Kelt migrated rapidly upon reaching the estuary and bay, swimming between 30 and 50 cm.s-1. It appeared that extended feeding within the delta permitted kelt to improve their physical condition, enabling them to migrate more rapidly through the marine habitat.
Another study mobilized passive tags at a micro scale to track the movements of smolt (younger smaller fish) in their fluvial habitat. The team resolved open scientific questions about smolts’ navigation capacity and their ability to sense salinity and the location of the ocean. Arrays of antennae in the stream bed enhanced spatial and temporal resolution by orders of magnitude. A previously unknown “commuter” behaviour of salmonid juveniles was observed and verified in subsequent seasons. The existing paradigm of energy optimization was revised to deal with much greater distances that these tiny fish actually travel.
These unique observations emphasize as never before the formidable migratory skills of a species that is symbolic of Canada’s increasingly threatened biodiversity. (Project Final report, 2009)
CODIGEOSIM – Geosimulation Tools for Simulating Spatial – Temporal Spread Patterns and Evaluating Health Outcomes of Communicable Diseases
(Phase IV – PIV-05)
This project started in Phase III (2005) and continued through Phase IV (2012). An impressive array of complex, multi-scale models and simulation platforms were created over the course of project PIV-05, addressing several aspects of disease propagation and risk. The efforts of this team contributed to Canada’s response to the H1N1 outbreak in 2009. Models of airport traffic and human interaction were crucial for this disease, but for others, such as West Nile virus and Lyme disease, include various intermediary animal hosts. The following computer platforms were designed throughout this project: Zoonosis, and SenartMAGS.
The ZoonosisMAGS Platform is a new generic geosimulation platform to help non-geographers and non-mathematicians integrate spatial data and specify disease transmission models in a general and user-friendly framework for zoonoses. The team developed tools that represent large territories (that is, Southern parts of Quebec and Ontario), using a hierarchical spatial model so that users can associate spatial data (such as land cover, elevation, snow cover) and population data (number of individuals of each compartment of each involved species) to the cells of different levels. This platform was applied to West Nile virus and to Lyme disease.
In collaboration with Dr. Godard’s team (University Paris 8), the GEOIDE team developed tools to analyze the mobility of people in recreational areas (that is, the Forêt de Sénart near Paris) to assess risk in areas infected with tick populations. They also developed a web mapping tool to enable people fill in questionnaires and describe their visits in the forest using a map accessible on line. Team members are currently completing tools to simulate the visitors’ behaviour in the park and to relate typical behaviours to the visitors’ socio-economic profile (pattern characterization). (Project Final report, 2011)
Results for Industry
Automating Photogrammetric Processing and Data Fusion of Very High Resolution Satellite Imagery with LIDAR, iFSAR and Maps for Fast, Low-cost and Precise 3D Urban Mapping
(Phase II – MNG#TAO)
One of our investigators, Dr. Yun Zhang from Phase II, developed a breakthrough image-fusion technique, filed a US patent application, and completed technology licences to three world-leading companies, PCI (Canada), DigitalGlobe (US) and InterMap (Canada).
The fusion technology significantly increased PCI’s export potential and was one of the company’s best algorithms, generating great worldwide interest for using PCI Pansharp. This technology was fully integrated into DigitalGlobe’s production line to produce the world’s most advanced, pan-sharpened QuickBird images. The SAR-MS fusion technique developed in this project will benefit InterMap’s value-added production.
In addition, a new camera lens and chip technology developed by University of New Brunswick (UNB) start-up company, SceneSharp, hopes to deliver sharp colour images that will provide fast, accurate intelligence for the security industry. So promising is SceneSharp’s technology that the company won first place and $145,000 in the New Brunswick Innovation Foundation’s (NBIF) prestigious Breakthru award of 2011. Dr. Zhang and two UNB business students, Jordan deWinter and Pablo Alvarez, received the prize. The UNB PanSharp technology was originally intended for use in satellites, and UNB has licensed it for use by companies around the world, including Google Maps.
Not only does the SceneSharp software automatically detect motion, identify the 3D location, shape and size of the moving object and then transmit a high-quality image, it does all of that with a 50 percent reduction in data size. The additional advantage of the SceneSharp system is that it can be fully automated, providing unmanned surveillance that can be of real value for military, commercial or consumer applications.
In addition, Dr. Zhang received the first Giuseppe Inghilleri Award from the International Society of Photogrammetry and Remote Sensing (ISPRS) during ISPRS Conference in Melbourne that was held from August 25 to September 1, 2012. This is an award that will be given every four years to a person who has significantly enhanced the applications of photogrammetry, remote sensing, or spatial information sciences. Dr. Zhang received this award for his outstanding work in image fusion and automated pan-sharpening to improve the resolution of colour satellite imagery.
Three Dimensionalizing Surveillance Network
(Phase IV – PIV-17
This project started in Phase I, and received 13 years of funding from GEOIDE. It is GEOIDE’s longest-lasting project and had impressive output. It focussed on developing a general system for sensing, distributing, interpreting and visualizing the real‐time dynamics of urban life. The key to this technology is the integration of accurate 3D urban models with real‐time video data. This system provides a compelling understanding of the current dynamic life of a city, while protecting the privacy of its inhabitants. Object detection and motion tracking algorithms are used to find and follow people and cars in pan/tilt video camera data. A proprietary online calibration algorithm rectifies these data to a stored model of the 3D urban scene. Each person and car identified in video is then represented as an anonymous sprite in a 3D visualization of the urban scene. Observers can choose how they visualize the scene: in a user‐controlled walkthrough or flythrough, or from the point‐of‐view of one of the actual agents presently in the scene. (Project Final Report, 2011)
A successful unmanned air vehicle (UAV) survey was performed at York University with the Aeryon Scout UAV equipped with a video camera. Some of the results are under consideration by the software industry (GeoDigital International, 2G Robotics) and some internationally funded projects (University of Southampton, UK).
The topic of research is very important for applications such as urban planning, surveillance and design. Underlying technology such as image collection, processing, management and robotics were involved.
Ice Classification Using SAR Imagery to Support Canadian Ice Service Operations
(Phase IV – SSII-111)
GEOIDE Project PIV-SSII-111 set out to generate an effective algorithm for the binary classification of ice and open water for Canadian Ice Service (CIS) operations. This project has a high degree of intra-project networking among University of Waterloo, CIS and MacDonald, Dettwiler and Associates Ltd. (MDA).
MDA provided $50,000 cash and 21 scenes valued at $3800 per scene to support the project. CIS and MDA validated the prototype (RADARSAT-II owner). The University of Waterloo and MDA are collaborating to develop an operational CIS implementation to distinguish ice from open water without operator input.
The detailed maps generated are expected to be a necessity for navigation in hazardous waters (to reduce risk and save time and fuel) and to provide input into climate models. Other northern nations and oil companies are interested in purchasing maps derived from the proposed algorithm. Such research in computer vision required to solve this problem has been and will continue to be used in other applications, such as medical imaging and video interpretation. (Project Final Report, 2011)
Results for Informed Decisions
Tsunami Loss Estimation and Emergency Planning
(Phase III – SII-59)
The West Coast of North America faces tsunami hazards from a number of far- and near-field sources; however, the most extreme danger is posed by the near-field Cascadia Subduction Zone (CSZ) which can generate great (magnitude 8 or greater) earthquakes.
After the December 2004 Indian Ocean tsunami, the District of Ucluelet on Vancouver Island asked for assistance to prepare for such an event. The province of British Columbia and the Government of Canada initiated the Tsunami Integrated Preparedness (TIP) program to help Vancouver Island communities develop enhanced response plans, install and upgrade communications and warning systems, post evacuation signage, and identify safe havens. Before this project, the community of Ucluelet had already developed an understanding of the CSZ tsunami hazard, established a response plan, and performed periodic warning exercises; however, it did not fully understand its vulnerabilities and potential losses within each hazard zone.
This project provided site-specific hazard and evacuation maps, estimates of vulnerabilities and potential losses, and simulations of the hazard scenarios and evacuation scenarios. The expanded set of candidate safe havens will make more evacuation pathways available, shorten the average distance to safety, and provide a much larger receiving area for the evacuating population. Tsunami preparedness brochures have been distributed and tsunami signs have been installed, based in part on the outputs of the tsunami wave propagation/run-up model. An important activity in this process was the presentation of hazard and evacuation simulations at a public information session to show how an emergency might unfold. This helped citizens develop a better understanding of the hazard and of the need to take evacuation planning more seriously.
GIS and simulation models played an important role in this integrated approach; and this study demonstrated how investments by federal, provincial, and local governments in GIS-based datasets can be leveraged to support hazard assessment and mitigation planning needs at the community level. One of the most interesting findings is that evacuees do not need to move a large horizontal distance to gain a safe elevation above the incoming floodwaters because of the relatively steep local topography. The project team coined the phrase “Ucluelet is blessed with elevation” to describe this positive mitigation factor. (Project Final Report, 2009)
Local Climate Change Visioning – Tools and Process for Community Decision-making
(Phase IV – PIV-32)
The project had substantive impact on policy, including contributing to Kimberley’s policy 75 adaptation policy recommendations, several of which are now being implemented, and the development of comprehensive policy recommendations for Delta’s Council on the challenges facing the community due to rising sea levels. Delta also adopted new planning terminology, including consideration of “radical” adaptation concepts such as “Managed Retreat,” indicative of new approaches to climate change. In Toronto, Project 32 resulted in on-the-ground design and construction projects to offset the effects of urban heat near Pearson Airport, as well as the use of geomatics (solar web-mapping) within on-going neighbourhood design processes.
In Nunavut, the Deputy Minister for Community and Government Services told the project team “The GN appreciates the work you folks are contributing to advancing the planning for communities. This is important work providing data that can be used for better planning results.” Alberta Environment is considering expanding the hybrid watershed modelling to multiple watersheds across the province. Finally, the project team was influential in the global climate change field by: providing precursors to the new form of integrated scenarios being advanced by the IPCC in its 5th Assessment; being invited to present to the US National Climate Assessment team workshop; and contributing to the intellectual debate on how to develop socio-economic scenarios through organizing a AAAS conference symposium featuring four IPCC co-authors/Nobel prize winners.
TrafficPulse: A Participatory Mobile Urban Sensor Web for Intelligent Green Transportation
(Phase IV – SII-PIV-89)
The City of Calgary identified transportation mode detection as a priority to better manage city traffic. TrafficPulse APP collects users’ trajectory points and data from smartphone accelerometers and magnetometers. The team proposed a new methodology using neural network-based artificial intelligence to identify the mode of transportation. The new method is based on the patterns of the distinct physical profile of each mode that consists of speed, acceleration, number of satellites in view and electromagnetic levels.
Private sectors have started to realize the power of location-based social networks for traffic applications. For example, a start-up company in Calgary is developing a system, called Transit Hub Calgary that is very similar to the TrafficPulse system. Dr. Liang, the project leader, met with the company and shared TrafficPulse’s experience. Future collaboration between the TrafficPulse team and the start-up company is possible.
Ice Classification Using SAR Imagery to Support Canadian Ice Service Operations
(Phase IV – SSII-111)
MAGIC is a software system designed and built by Dr. Clausi (project leader) with his research team at the University of Waterloo. MAGIC can be used to segment and classify digital imagery with an emphasis on full-scene remote sensing imagery. The key scene classification algorithm incorporated within MAGIC is called IRGS. Initially, the primary goal of this project was to use IRGS for identifying all ice types and open water using coarse ice maps provided by ice analyst. When the project started, CIS (Canadian Ice Service) recommended solving a different but related problem; namely, the binary classification of ice and open water without any ice analyst provided maps. This changed the focus of the research to produce an operationally preferred algorithm.
The implementation of an ice/water classification system at Canadian Ice Service (CIS) will have the following benefits:
- Improved accuracy of overall ice and ice type concentrations. The proposed ice/water classification system can accurately calculate ice concentrations. When these data are fed into ice and climate models, more trustworthy predictions can be made.
- Improved efficiency of CIS operations. An ice/water classification system would relieve the operator of assessing ice concentration; improve the processing time and reduce analyst’s work load.
Automated production of sensor-resolution ice maps will provide commercial/social benefits such as improved ship navigation, improved route-planning for ice breakers, improved understanding of regional climate processes, improved data-quality for end users, and more commercial opportunities for the Canadian high tech industry.
Results for Highly Qualified Personnel
The GEOIDE Students’ Network and the GEOIDE Summer School
GEOIDE students are in many respects the principal result of the Network’s operations. Their exceptional training was due to the interdisciplinary framework of projects and the partnerships with all sectors concerned. Over its existence, the GEOIDE Network has contributed to the training of about 1400 students that now compose a significant part of the new generation of geomatics professionals and scientists working in Canada and abroad.
Since its beginning, the GEOIDE Network has encouraged and supported two major student’s initiatives: the GEOIDE Students’ Network (GSN) and the GEOIDE Summer School (GSS). The GSN and GSS allowed GEOIDE students to see beyond their specific research projects and gain a more complete academic experience and professional training through collaborations with large interdisciplinary body of students. It has helped students’ transition from an academic environment valuing relationships with their supervisor and other students, to a professional environment valuing relationship with their peers that can benefit their entire professional life. While a number of GEOIDE students decided not to take this opportunity, those that did have realized that in such network, “the whole is greater than the sum of its parts”.
Perhaps one of the greatest successes for GEOIDE and the GSN/GSS was to enable a culture of collaboration amongst a new generation of geomatics professionals and scientists that came from very different backgrounds and cultures. Many of the students who engaged in the GSN now have careers that emphasize collaboration and multi-disciplinary work; collaboration comes naturally as they plan their projects. Additionally, the Canadian geomatics community is now much more connected than it was before GEOIDE, as most the 1400 HQP trained under GEOIDE now have geomatics-related jobs. Students trained at different universities, such as the authors of this paper, were connected through GEOIDE and are now geomatics colleagues initiating new pan-Canadian collaborations. Having a network of colleagues has supported GEOIDE graduates in early career stages by providing opportunities to seek advice, share students, and conduct research collaboratively.
The success of the GSN and GSS is to our knowledge unique amongst Canadian NCE networks. It has inspired other networks and has played an important role in the review GEOIDE received over the years. While the concept of a student network was a strong point in the initial proposal, its success has been positively received by the expert panels assessing the different GEOIDE funding renewals. This in term translated into a constant support from GEOIDE for student’s initiatives, which turned to be a win-win situation for both.
See chapter 2 of The Added Value of Scientific Networking: Perspectives from the GEOIDE Network Members, 1998-2012 for the complete story of the GSN and the GSS initiatives.