Between the Poles

Key initiatives of importance to the global geospatial+AEC subsurface digital twin community

Two initiatives of importance to the global geospatial community will begin this June.

Expanding Responsibility for the BetweenThePoles blog

First of all we are expanding the BetweenThePoles blog from a single content provider (Geoff Zeiss) to an industry collective that will be able to mine many more geospatial industry players. Geoff is stepping down from day to day responsibility for the blog (he doesn't like the word retiring) and ownership and administration will transition to GoGeomatics Canada, advised by a collective of industry leaders in the geospatial+AEC sector.

The content and aims of BetweenThePoles will remain the same. The blog encompasses the application of geospatial technology in architecture, engineering, and construction (AEC) and utilities. Primary focus areas are worldwide legislation, governance, regulations, and new technologies for mapping underground utilities and enabling subsurface digital twins, CAD+GIS and BIM+geospatial integration to enable full lifecycle construction, and innovative technologies for vegetation management for electric power utilities.

By way of background the BetweenThePoles blog was initiated in 2006 by Geoff Zeiss. Since then the blog has hosted over 2,500 posts and received over 1.3 million lifetime page views. Currently 3,500 active users access the blog each month. The blog will be continued with its current focus; mapping underground infrastructure,  BIM+geospatial integration, and innovations in vegetation management for utility transmission and distribution.

Making Subsurface Utility Mapping Strategy Forum (SUMSF) and Annual Event

The second initiative follows on the very successful Subsurface Utility Mapping Strategy Forum (SUMSF 2022), which was a follow-up event to the Canadian Underground Forum in 2021.

The genesis of SUMSF is found within Canada’s National Geospatial Conference, GeoIgnite. In 2021 Jonathan Murphy and Geoff Zeiss came together over the idea to create a leadership and technology forum on the theme of underground infrastructure mapping. With the help and financial support of the global underground mapping industry the Canadian Underground Forum (CUF 2020) was held virtually in 2021. From this success, a new and larger stand-alone event with a new global brand and perspective was launched in 2022, the Subsurface Utility Mapping Strategy Forum (SUMSF 2022). 

SUMSF 2022 was held virtually March 22nd and 23rd 2022 again with the support of the underground mapping industry. SUMSF was themed to bring together individuals and organizations to share information about and contribute to improving the mapping of underground infrastructure. This online conference brought together exceptional speakers in the underground locating and mapping sector and was well attended by the geospatial and engineering community.

Based on the success of these two events GoGeomatics Canada has decided to make the conference an annual event. Next year, SUMSF will be held virtually March 6th and 7th. SUMSF is hosted and presented by GoGeomatics. Like BetweenThePoles blog, it will be advised and supported by industry leaders with the support of the SUMSF LinkedIn Group. 

BetweenThePoles will remain an important platform for communication and engagement on the the application of geospatial technology.  

Anyone interested in joining the collective for the BetweenThePoles blog are encouraged to contact Jonathan Murphy at GoGeomatics Canada. As well anyone interested in joining the advisory committee for SUMSF 2023 is encouraged to do the same.

Geoff Zeiss

Real time ground deformation monitoring from satellites to reduce risk of damage to utilities and pipelines

Several years ago I blogged previously about an application that used Sentinel-1 radar interferometry to monitor ground displacements with millimetre precision to look for subsidence which might be evidence of a leak in a municipal water system. Now access to data and processing algorithms are available to developers to create workflows to monitor ground deformation to reduce the risk of damage to underground Infrastructure including utilities, oil, gas and water transmission pipelines and associated industrial plants.

Ground movement represents a significant risk to linear structures such as utility, pipeline, and transportation networks as well as to large scale industrial and commercial structures.  Traditionally ground deformation is monitored by fixed high accuracy GNSS stations located at intervals and by periodic surveys along alignments.  Both of these are expensive and are limited by geospatial and temporal gaps. Satellites such as Sentinel-1 that use radar to image the earth are capable of radar interferometry (INSAR) which allows ground deformations on the order of cm and even mm to be detected.  The revisit times for these satellites are on the order of days and the radar antennas on these satellites allow very large areas to be monitored for deformations.  For example, Sentinel-1 is capable of detecting deformations of a few mm over an area of 42,500 km² with a 12 day revisit time. Furthermore INSAR data is available stretching back to the mid 1990's making it possible to monitor historical ground movement over quite long time periods.

UP42 provides the infrastructure of an enterprise software development and deployment platform enabling developers to create scale-able workflows using satellite imagery and customized algorithms. In a recent UP42 webinar (53 minutes in), Shawn Melamed and Sarah Thompson of CATALYST (PCI Geomatics) showed how innovative INSAR technology has made it possible to cost effectively continuously monitor ground deformation along thousands of km of  a linear structure such as a utility, pipeline, or transportation network.  The data can be sourced from any of the satellites capable of INSAR including Sentinel-1 (free) and TerraSAR-X (commercial).  An example is provided in the talk of the development of a UP42 workflow for detecting uneven land subsidence in California resulting from agricultural withdrawals of water from an aquifer that potentially could damage a power plant located in the area.  The code to perform this analysis is available as the Ground Displacement Block on the UP42 marketplace.

 

 

Breakthrough broadens adoption of GPR in 3D mapping of underground utilities

Traditional output from a GPR scanner showing hyperbolas reflected from underground utilities

The major inhibitor to broader use of ground penetrating radar for detecting the location of underground utilities is the difficulty in interpreting radar scans.  The Technology Report Technology Advancements and Gaps in Underground Safety published by the Common Ground Alliance in 2018 identified "User-friendly GPR system to deploy for helping construction crews avoid utilities" as an important gap.  In September 2019 at Intergeo Leica Geosystems released the DSX ground penetrating radar system that is the first to address this gap in a compact, portable system.  The DSX system includes analytic software DXplore enabling GPR to be used productively without a trained geotechnics specialist by utility repair and maintenance, civil engineering and surveying companies.

Before the DSX, non-geotechnical professionals including surveyors found it a challenge to interpret GPR scans consisting of images with hyperbolas showing the reflections of RF waves from underground objects.  GPR remained the preserve of highly trained geotechnical and other underground engineering professionals.

As I discovered at a recent talk I gave to the Alberta Land Surveyors Association, there is a lot interest among the surveying community to expand their professional offerings beyond the traditional above ground survey to include below ground surveys.  Two years ago the one-button BLK360 opened up engineering grade laser scanning to a much broader professional audience.  Leica Geosystems is betting that with a similar focus on simplicity the DSX device and DXplore software is going to open GPR scanning to a much broader professional community, including civil engineering and surveying companies.

Output from the DSX with DXplore software showing tomographic image of detected underground utilities

The key feature of the DSX GPR device and DXplore software is analytics to interpret the GPR scans and visualize underground infrastructure in the form of tomographic images, rather than radar hyperbolas.

For accurate locating the DSX supports seamless integration of positioning sensors (GNSS antennas) and is the first device to be integrated into machine control workflow through avoidance zones based on an underground utility map generated by the DSX.

The DSX workflow is very simple: import available as-builts in the area as DXF or other CAD format files into DXplore.  If possible, walk the site to perform a visual inspection using an electromagnetic locator (EML) to search for power cables and conductive utilities and mark them on the ground.  Define a grid for the area to be surveyed.  Walk the grid with the DSX device linked with Leica GPS/GNSS antennas with real-time feedback on the DSX's display.  Create a georeferenced 3D map using the tomographic display (no hyperbolas !) of what was detected to manually identify suspected utilities on the display.  The software then processes the captured GPR images to confirm the utility pipe or cable.  You can then export a 2D or 3D vector file in a CAD compatible file format.  Very simple.

But too simple for experienced GPR practitioners. I have talked to very experienced GPR people who don't find the DSX interesting for them. As GPR experts they are trained to work with B-scan/radargram data representation and interpretation, thus expert systems such as Leica DS2000 or IDS GeoRadar Stream are more suitable.

But for people who don't have GPR experience, believe that underground surveying represents a significant business opportunity, and already use or know Leica Geosystems total stations or other survey equipment this is being seen as a way to get their toes in the water.  Furthermore I fully expect that Leica will not stop here but will continue to develop easy to use GPR software for detecting and mapping underground infrastructure including support for the multi-channel Leica IDS Georadar arrays that are being applied to very sophisticated mapping of underground infrastructure.

New regulations for pipeline safety creates opportunities for geospatial professionals

New federal regulations in the U.S. will likely more than double the areas where higher levels of pipeline integrity maintenance standards are mandated by federal legislation.  Identifying these areas is more complex creating opportunities for geospatial professionals to apply new technologies including machine learning.

On June 10, 1999, a gasoline pipeline operated by Olympic Pipeline Company exploded in Bellingham, Washington. Three people died in the accident.  On August 19, 2000 a natural gas pipeline owned by the El Paso Corporation exploded near Carlsbad, New Mexico killing 12 people.  After these incidents the government moved quickly to improve the quality of geospatial and other data about underground pipeline assets. The result was within about two years federal regulation from the Pipeline and Hazardous Materials Safety Administration (PHMSA) which required companies operating pipelines carrying hazardous materials to identify high consequence areas (HCAs) where there were 20 or more structures intended for human occupancy within a radius (potential impact radius or PIR) defined by the diameter and pressure of the pipe. Within the HCAs utility companies were required to maintain facilities data at a higher level of integrity. The cost to identify HCAs for a relatively small utility in Pennsylvania that I was familiar with at the time cost $2 million over two years.  This represented a significant opportunity for experienced geospatial professionals.

The Pipeline Safety, Regulatory Certainty, and Job Creation Act of 2011 mandated telephonic or electronic reporting of accidents and incidents including the amount of product released, number of fatalities and injuries, but not an accurate location, within 48 hours of occurrence.  The act also required the Secretary of Transportation to maintain a map of high-consequence areas in which pipelines are required to meet integrity management program regulations.  At this time the locational accuracy requirement for pipelines was  ± 500 feet.

At the GITA Pipeline Forum in Houston, Darrel Donaho of G2 Integrated Solutions presented an overview of the latest regulations from PHMSA.  Referred to as the Gas Mega Rule the new regulation extends the higher standards of pipeline maintenance to moderate consequence areas (MCAs). Pipeline operators are now required to identify areas where 1) the number of occupied structures within the PIR is five or greater and less than 20 or 2) where any portion of the pipeline PIR intersects the paved area of interstate highways, freeways and expressways and other principal four-lane arterial roads.  They are also required to report longitude/latitude to five decimal places for all incidents.

Whereas in the past identifying HCAs involved hand drawing building footprints from imagery, machine learning can now be used to automate the process of identifying building footprints prior to using a buffer analysis to identify areas qualifying as MCAs.  The analysis is aided by the current accuracy requirement for locating pipelines of ± 50 feet.  Identifying MCAs where highways intersect pipeline PIRs is more complex because automating the identification of transportation corridors is more difficult.  There are alternative, but more expensive ways to find this information.

An analysis of a 14.91 mi transmission line suggests that the impact of the new regulation will be significant, more than doubling the areas where higher integrity maintenance applies.  This creates a significant opportunity for geospatial professionals to apply new technologies including machine learning.

HCA	MCA (>=5)	MCA roads	Total MCA
5.65 mi	2.63 mi	3.03 mi	5.66 mi

 

Applying machine learning to identifying housing starts from satellite imagery

Deep learning is increasingly finding practical applications using geospatial data, specifically, satellite imagery. At the GeoIgnite conference in Ottawa, Nicolas Martinez of Statistics Canada described a project scheduled to kick-off in July to use machine learning to identify housing starts across Canada.  The goal of the project is to improve the accuracy and coverage of housing starts survey, specifically to fill data gaps for smaller remote and indiginous communities which are generally excluded from the current survey program. 

Canada Mortgage and Housing Corporation (CMHC) currently employs a significant number of people at their head office in addition to field agents across the country who physically confirm residential starts and completions.  The objective of the project is to apply convolutional neural network (CNN) machine learning to satellite imagery to extract information about housing starts and completions. The imagery will be preprocessed to prepare it to be used in developing a machine learning processing model.  The largest challenge will be developing training and testing imagery data sets.

The original application of deep learning to photo imagery was by Geoffrey Hinton at Google who used it to distinguish dogs and other objects in photos that people uploaded.  More recent applications of satellite imagery targeted landuse and agricultural identification. For example, remote sensed imagery from a satellite can be used to create a model that will differentiate between corn and potatoes, using factors that can be calculated from satellite imagery such as normalized difference vegetation index (NDVI) (min, max, mean), texture (min, max, mean), vegetation height, and geometric factors such as orientation.  I blogged about how open source code and publicly available training data has been applied to track deforestation and reforestation in Mato Grosso, a state in the central amazon region of Brazil. A study in the Netherlands used CNN to identify blocked waterways using overflight imagery with a success rate of 97 %.

An important application of satellite imagery is identifying built structures. I have blogged about how satellite imagery has been used to identify buildings and transportation networks by applying a neural network model together with OpenStreetMap layers and high resolution Worldview multispectral imagery.  NVIDIA has demonstrated the ability to automate detection of many road networks using deep learning algorithms and multi-spectral high resolution imagery.  Another application of CNN and imagery in the Netherlands was able to differentiate residential roofs with dormer windows from those without.

The essential key to effective application of deep learning is good training and test data - ground-truthed data that involves, for example, someone on the ground identifying whether the fields seen in the imagery are corn or potatoes or something else. There are  publicly available training data sets that can be used to train satellite imagery CNN models. For example, I blogged about the System for Terrrestrial Ecosystem Parameterization (STEP) dataset which has 2000 manually labeled sites covering 17 different land cover types scattered across all continents. A large publicly available training dataset derived from Sentinel 2 imagery contains 30,000 polygons of land use training data for ten classes of land use: annual crop, forest, herbaceous vegetation, highway, industrial, pastures, permanent crop, residential, river and sea or lake.

The Statistics Canada project is scheduled to begin in July with initial work on a proof of concept and CNN testing to begin in August and September.  The communities chosen for the initial phase are  Kitchener-Cambridge-Waterloo, Red Deer, and Iqaliut.  The first phase is scheduled to be completed by winter 2019.  The project plans to use Tensorflow and Keras, open source libraries and a python framework for applying machine learning.  Artificial intelligence (AI) has often suffered from inflated expectations, complex code, heavy processing requirements and not very practical applications. There is a growing body of practical applications in the geospatial domain that shows that as a result of the immense processing power available in the cloud machine learning is able to generate fairly easily practical and useful results.  As Chris Holmes pointed out in his talk at FOSS4GNA, the challenge for Statistics Canda will be to develop reliable and publicly available training and test datasets to enable deep learning models to be created for a broader range of applications using the huge volumes of satellite imagery and other geospatial data that are now available.

Deep learning is increasingly finding practical applications using geospatial data, specifically, satellite imagery. At the GeoIgnite conference in Ottawa, Nicolas Martinez of Statistics Canada described a project scheduled to kick-off in July to use machine learning to identify housing starts across Canada.  The goal of the project is to improve the accuracy and coverage of housing starts survey, specifically to fill data gaps for smaller remote and indiginous communities which are generally excluded from the current survey program. 

Canada Mortgage and Housing Corporation (CMHC) currently employs a significant number of people at their head office in addition to field agents across the country who physically confirm residential starts and completions.  The objective of the project is to apply convolutional neural network (CNN) machine learning to satellite imagery to extract information about housing starts and completions. The imagery will be preprocessed to prepare it to be used in developing a machine learning processing model.  The largest challenge will be developing training and testing imagery data sets.

The original application of deep learning to photo imagery was by Geoffrey Hinton at Google who used it to distinguish dogs and other objects in photos that people uploaded.  More recent applications of satellite imagery targeted landuse and agricultural identification. For example, remote sensed imagery from a satellite can be used to create a model that will differentiate between corn and potatoes, using factors that can be calculated from satellite imagery such as normalized difference vegetation index (NDVI) (min, max, mean), texture (min, max, mean), vegetation height, and geometric factors such as orientation.  I blogged about how open source code and publicly available training data has been applied to track deforestation and reforestation in Mato Grosso, a state in the central amazon region of Brazil. A study in the Netherlands used CNN to identify blocked waterways using overflight imagery with a success rate of 97 %.

An important application of satellite imagery is identifying built structures. I have blogged about how satellite imagery has been used to identify buildings and transportation networks by applying a neural network model together with OpenStreetMap layers and high resolution Worldview multispectral imagery.  NVIDIA has demonstrated the ability to automate detection of many road networks using deep learning algorithms and multi-spectral high resolution imagery.  Another application of CNN and imagery in the Netherlands was able to differentiate residential roofs with dormer windows from those without.

The essential key to effective application of deep learning is good training and test data - ground-truthed data that involves, for example, someone on the ground identifying whether the fields seen in the imagery are corn or potatoes or something else. There are  publicly available training data sets that can be used to train satellite imagery CNN models. For example, I blogged about the System for Terrrestrial Ecosystem Parameterization (STEP) dataset which has 2000 manually labeled sites covering 17 different land cover types scattered across all continents. A large publicly available training dataset derived from Sentinel 2 imagery contains 30,000 polygons of land use training data for ten classes of land use: annual crop, forest, herbaceous vegetation, highway, industrial, pastures, permanent crop, residential, river and sea or lake.

The Statistics Canada project is scheduled to begin in July with initial work on a proof of concept and CNN testing to begin in August and September.  The communities chosen for the initial phase are  Kitchener-Cambridge-Waterloo, Red Deer, and Iqaliut.  The first phase is scheduled to be completed by winter 2019.  The project plans to use Tensorflow and Keras, open source libraries and a python framework for applying machine learning.  Artificial intelligence (AI) has often suffered from inflated expectations, complex code, heavy processing requirements and not very practical applications. There is a growing body of practical applications in the geospatial domain that shows that as a result of the immense processing power available in the cloud machine learning is able to generate fairly easily practical and useful results.  As Chris Holmes pointed out in his talk at FOSS4GNA, the challenge for Statistics Canda will be to develop reliable and publicly available training and test datasets to enable deep learning models to be created for a broader range of applications using the huge volumes of satellite imagery and other geospatial data that are now available.

A killer application for satellite imagery collected daily for the entire earth

At the GeoIgnite conference in Ottawa, Wade Larsen of Urthecast described what may be a killer application for satellite imagery captured daily. Farmers in the US over-irrigate by 25% per year. A recent study has demonstrated that monitoring vegetation growth on a daily basis can save farmers $50 000 in water per year. With 260,000 pivots operating in the US, this amounts to billions of dollars of savings every year.

Urthecast has been known for having two cameras (Iris and Theia) on the International Space Station and two Deimos earth observation satellites. Last year Urthecast announced the acquisition of Geosys Technology, a software company that processes earth observation imagery for agricultural applications.  Their latest initiative is called UrtheDaily with the objective of capturing nine bands of imagery from six satellites covering the Earth's entire landmass every day. The nine bands are identical to Sentinel-2's visual and near-infrared bands. The cameras are cross-calibrated with Sentinel-2 to assure data compatibility. The six cameras enable a daily revisit for higher latitudes and somewhat less for tropical areas. It captures imagery at the same time, from the same altitude every day.  About 150 terabytes will be acquired and downloaded every day. The imagery has five meter resolution (GSD) and represents scientific grade data quality that will be available on the cloud for user access within hours of acquisition. UrtheDaily is designed to support machine-learning and AI ready geo-analytics applications on a global scale. The UrtheDaily platform will provide the ability to discover data, run user algorithms, and deliver data to customers.

UrtheCast reports that it has already signed contracts primarily in the agriculture sector. An example of a potential killer app in agriculture is monitoring irrigation. Farmers in the U.S. overwater by 25 %. A four year in Texas by HydroBio which assessed the value of frequent monitoring of plant health and growth using satellite data concluded that to be effective this required high-frequency (daily) monitoring with highly calibrated sensors. With its resolution UrtheDaily will be able to monitor vegetation for individual pivots on a daily basis in order to optimize water utilization. It is estimated that this could save farmers $50,000 per pivot per year. (A pivot typically irrigates 120 acres.) There are 260,000 pivot irrigation systems in the U.S. which translates into the potential to save farmers billions of dollars per year.

To date UrtheCast reports that it has signed $350 million worth of contracts in the U.S., Russia, India and other countries. It plans to launch six satellites in 2021 and become operational in 2022.

The future is fewer maps

At the first Longitude meetup in the U.K., organized by Ben Flanagan, Kenneth Field, and Chris Wesson, Ed Parsons presented his perspective on one of the directions to expect from geospatial in the future.  It was a bit counterintuitive, for consumers more geospatial data and technology, but fewer maps.

To start with he reminded us that the public has never had as much access to maps as we do now thanks primarily to Google Maps - either as standalone or embedded in other apps. What he foresees is that in future we will be seeing fewer maps or no maps at all.

The maps we have used up to now are large scale maps presenting a large scale view - probably very useful to Winston Churchill and people planning things on a large scale. But in our day to day life, we don't need maps at this scale. Instead we need to know what we will encounter in the next 15 minutes, whether that be the next exit off the highway or the next grocery store. There will be more interaction between us and the world around us - for example showing which of the nearby bars or coffee shops are very busy or less busy.

Ed gave a practical example that anyone who uses a smart phone to navigate in a place they have not been before can relate to. When you come out of a tube station or any GNSS-denied place you are always faced with a quandary - how to orient yourself to know whether to walk left or right. I mentally flip a coin and walk until my smart phone tells me if I am headed in the direction I want to go. If I am in a part of the city where GNSS is unreliable I walk to the end of the block and the street number or name usually will tell me if I am headed in the right direction or that I have to retrace my steps in the other direction.  In the future your navigation app won't rely solely on GNSS signals but will recognize buildings and give you the right direction. In fact this could all happen through an audible interface so that you needn't look at a map at all, "turn right as you leave exit 3 from the tube station".  Of course this will require a lot more spatial data about the city you are in and will require some serious geoanalytics.

What this suggests is that as consumers we will use more geospatial data and analytics, but it will become part of the technology stack and most of us won't realize how much geospatial technology we are using.

Geospatial data and analytics applied to things you never imagined

Orbital Insights, a company headquartered in the World Trade Center Tower 1, accesses data generated by satellites, drones, balloons, smart phones, and sources which it process and analyze at petabyte scale to make it actionable for businesses, governments, and NGOs. Last evening at a reception for people attending GEOGRAPHY 2050 Powering Our Future Planet, I had an opportunity to chat with a couple of their technical experts about some the ways they are applying geospatial data that I doubt anyone would have imagined just a few years ago.  As an example of a commercial application Orbital Insights tracks global oil Inventories which lets you follow oil stock fluctuations on a map.  Their data correlates with the U.S. Energy Information Administration (EIA) data, but is available 2-3 days before EIA data is released which provide early valuable insight into price fluctuations.

Counting cars in airport parking lots as a measure of economic activity

Using satellite imagery from Digital Globe and other providers, Orbital Insights used machine learning to identify and count the number of cars in parking lots at Damascus and Aleppo airports.  One of the periods they analyzed was during and after the government's successful attack on Aleppo.  While the number of cars parked at Damascus airport remained fairly constant, at the cars parked at Aleppo's airport gradually dropped as the government attack intensified.  Ultimately it dropped to zero and then gradually began to increase again.  But inspecting the high resolution imagery, it was possible to make out that the aircraft parked at the airport after the fall of Aleppo were not commercial as before, but military aircraft.

Monitoring cell phone activity as a measure of car manufacturing activity

Anonymized cell phone activity data, which is available commercially from Apple, Google and others, carries a location and time stamp.  In a creative application of how this data cal be used, analysts at Orbital Insights defined polygons that enclosed the automobile assembly plants of two car manufacturers in the southern U.S. This provided a surrogate for activity on these assembly lines.  They looked at a period that included a couple of weeks prior to a major hurricane coming through the area and for week or two subsequent to the hurricane.  They were able to determine that one of the manufacturers ramped production up sharply before the hurricane hit while the other proceded in a business as usual mode.  They were also able to determine how long it took to restore production to pre-hurricane levels.

In both cases it is difficult to imagine how this kind of information could have been acquired and made public using a traditional approach.  I doubt that the Syrian government would have been forthcoming with this data, and in the second case this is competitive information that would have been difficult to acquire and release publicly.

Using publicly available environmental data in a spatial analysis of low birth weight babies

A spatial analysis of possible environmental factors contributing to the increase in the number of underweight babies born in Canada was presented by Charlene Nielsen at GeoAlberta 2018. Termed low birth weight at term (LBWT), this is the second most important cause of infant mortality in Canada and costs the medical systems hundreds of millions of dollars every year. It is also linked to childhood and adult diseases. Charlene's hypothesis was that environmental factors including chemical pollution, the man made environment, and socioeconomic level were contributing factors to low birth weight babies in Canada.  She used spatial analytics to study the problem which she used as the basis for her PhD thesis at the University of Alberta.

Using open, publicly available databases Charlene related the spatial distribution of low birth weight babies to geographical proximity to air-based chemical emissions or to land-based pollutants related to electrical power lines, dumps and solid waste sites, gas stations, mine sites and tailings, oil and gas well pads, transformer stations, and crop lands.

One of the publicly available data sources for chemical emissions that she used is the National Pollutant Release Inventory which has been maintained by Environment and Climate Change Canada since 1993.  It currently monitors the emissions of 324 chemical substances.

Her study found that air-released or land-based pollutants may be more important depending on geographical location. There was a greater correlation of low birth weight babies with the environmental factors she studied in Ontario, Quebec, Alberta and British Columbia. While this study found that  environmental factors did contribute to low birth weight babies, these are not the only factors. This study was particularly interesting because it used the geographical distribution of low weight births and the spatial distribution of environmental factors to try to determine some of the factors contributing to low weight births. Secondly it used publicly available environmental data which demonstrates that this type of spatial analysis can be easily applied in other studies.

Enabling the rest of us to find actionable information in huge volumes of satellite data

There is a huge volume of satellite imagery that is being captured and has been captured over the past couple of decades.  The challenge is to automatically find objects such as ships, trees, buildings, or containers in this vast amount if data and track changes in them over time.  This morning at at this year's FOSS4GNA (Free and Open Source Software for Geospatial North America) get together in St Louis, Chris Holmes of Planet Labs, gave an insightful overview of the application of deep learning to geospatial data, identifying opportunities for the open source community to lead in the application of this data and technology to addressing this challenge.

Outside of Google, Planet Labs, DigitalGlobe and some others deep learning has been applied primarily to recognizing people in selfies and dogs and cats from peoples photos on online repositories such as Flickr. I have blogged previously about competitions DigitalGlobe (which has 100 petabytes of imagery) and Nvidia have hosted into automating building footprint and road network recognition. Planet Labs has a computer vision and analytics team that has been applying the results of academic research to geospatial data, primarily imagery from Planet's constellation of 150+ satellites. For example, they have used deep learning to detect building footprints and road networks in Dar es Salaam. Chris showed examples of airplane, ship and container recognition achieved using deep learning. But these are research pilots. One company has applied deep learning to scale for a practical application. Google has been developing and applying the technology to building recognition - automatically identifying building footprints and features such as aerials. The deep learning algorithms have been developed by academia and as a result the code for the most part is open source.  For example, a  deep neural network model developed originally for medical image segmentation called U-Net  has been applied to identifying building footprints.  Successful training of deep networks requires many thousand labeled training samples and this training data is not open source. Labeled data involves people on the ground manually ground-truthing building footprints and other features so that the deep learning algorithms can learn what to recognize.

Chris believed that it is time to open these models to a broader group. There are some technical things that need to happen to make this possible. We need new formats for exchanging raster and vector data. Cloud optimized GeoTiff (COG) makes it possible to stream imagery from cloud sources such as Amazon S3. Planet, Google Earth, and QGIS already support COGs and there are other open source tools such as COG-Explorer that support them as well. Chris and others are working on WFS 3.0, a completely new developer-friendly version of an existing OGC (Open Geospatial Consortium) standard for exchanging vector data over the web.

Another technical advance that is required is a common way to query satellite data based on date, time, location, type of bands, and so on.  Chris and others have been working on a standard for searching satellite imagery for objects. The SpatioTemporal Asset Catalog, known as STAC, is an open specification that came about when fourteen different organizations came together to increase the interoperability of searching for satellite imagery. Currently when a user wants to search for all the imagery in their area and time of interest they can’t make just one search — they have to use different tools and connect to API’s that are similar but all slightly different. The STAC specification aims to make that much easier, by providing common metadata and API mechanics to search and access geospatial data.  Chris emphasized that standards are essential for opening up satellite data to processing and visualization and encouraged the open source community to get involved with the OGC to further developer-friendly standards.

Currently in order to use imagery data, a lot of data preparation is required of the end user such as lining up pixels, assigning a projection, removing clouds and mist, and other corrections. We need to move toward analysis ready data where data derived from different sources can be combined for analysis.

Because of the huge volumes of computational intensity we need cloud-native engines to perform the analyses. Google, DigitalGlobe (Maxar), Planet have these deep learning engines, but Chris encouraged the open source community to get more involved in developing these engines.

An essential piece that is required to open up the data and technology to a broader audience is open labeled data, Chris referred to it as a "open labeled data commons". There are analogies with OpenStreetMap initiative. Chris suggested that something like this is required for labeled data - it needs a repository of open data, tools to capture and manage the data, and a community around this to capture and curate that data.

The other key piece is converting this data into actionable information.  For example, dashboards that would alert municipal governments to un-permitted development, forestry enforcement to illegal forest cutting, or commodity price monitoring agencies to changes into the flow of ships into harbours or to large changes in containers at container ports.

Taken together Chris suggestions and work underway represent a roadmap for the open source geospatial community to lead in making the huge volumes of satellite imagery and other data queryable and able to generate actionable information.