Author: Andrew Iwanoczko
At a moment in time when emerging technologies are simultaneously being devoured by start-ups and innovators - I was fortunate enough to meet with Joe Bailey of Future Cities Catapult to share experiences on these what the future of our cities hold and how we might get there.
AI: Hi Joe. In your role as Data Scientist at Future Cities Catapult, I am guessing you see a range of technology advancement - or at least ambition and incremental steps toward that end.
As we see a rise in fragmented innovation that harnesses the power of, say, the Internet of Things (IoT) -these innovations are often termed as “Smart”. But in your view, aren’t these technology advances simply clever and useful innovation? Ought we not consider these innovations “Smart”,but for when such primary innovation is associated and inter-connected with something currently unrelated?
JB: Theoretically you could argue that smart things have been happening forever. For example, one could argue that even our surnames, sewer systems and traffic lights were all smart technologies in their time - each made a notable incremental difference in the functioning of society - arguably for the better.
As a result of this, the notion of Smart is just a spectrum of ‘useful’ and the smartness of a technology is directly reliant on its user. Something as simple as a straw that filters water is a smart device in some parts of the world. In others, it’s something like a self-driving car. Again, the value of these smart technologies is directly reliant on the society that they seek to improve. As a consumer, is a ‘smart’ thermostat of any value to me if I’m struggling to put food on the table?
Many suggest that the idea of two way communication makes something Smart. Others suggest that it’s this two way communication and an opportunity to adapt to a situation - i.e. the capacity for learning quickly which may add more value to a device or technology. I think it’s this capacity for quick learning that sets something apart at the moment.
One of the most important things to consider with smart devices however, is: who do they benefit? There is the potential for smart devices to exacerbate social inclusion - for example are they too expensive, can they be used by all and do they benefit society as a whole? There is certainly an argument that they maybe paving the way for other benefits later on that will address all, but being aware of the net societal value is key.
AI: You bring us onto a humanistic view of what is Smart based on the “benefit” that any piece of innovation has on a particular community. How much of the development of Smart-technology as cornerstones of Future Cities is in your opinion the responsibility of Government then? Is there any responsibility at all, and if so what steps are currently being made? And I suppose, based on where we are now, where do you see the potential of government to be doing more to support this in say 10 or 15 years time?
JB: Interestingly in Bill Gates’ most recent note on the ‘Gates Notes’he has said that “Innovation starts with government support for the research labs and universities working on new insights that entrepreneurs can turn into companies that change the world. The public sector’s investments unlock the private sector’s ingenuity.” As a starter for ten this suggests that someone (arguably very familiar with innovation) recognises that there is definitely a role for government in Innovation and by extension smart cities.
On a personal level I believe that one of the first roles of government is to create the policy environment that enables the open market to make life better in cities – this may be tax breaks for intelligent energy systems; policy changes for experimental testing; data releases for data analytics; tax breaks for innovation agencies; or something else.
This is of course complicated by the different levels of government but I believe that there are efforts to spur innovation around smart cities. The existence of the Future Cities Catapult -and note the Future not Smart - is a result of direct government action. We exist to help correct the market failure surrounding the future of our cities. We also see events like the government initiated smart city forum and evidence of other interventions from the government. The influence of government is also very sector specific and different sectors are more likely to be involved than others. Infrastructure is a good example.
Fortunately, private industry also recognises where it has an opportunity to contribute and it will do - particularly when there is an economic argument for contribution. We already see evidence of this and often,governments already buy city products from the market. In short, without government support it will be very difficult - cities are civic constructs and a mix of people, spaces,businesses and public sector stuff - these are often inseparable.
AI: In an article by McKinsey & Company, “5 technologies for the next 10 years”, the writer considers an elaborate interconnection between mobile devices, internet of things, machine learning, robotics and block chain technologies. The article speaks of “a leaky well-head gasket” in the oil and gas domain, but it could be anything mechanical that needs fixing.
In it, a machine learning algorithm detects a potential problem in a well-head and based on trained examples throws up a support incident related to what is very likely to be a leaky gasket.
Next we see automated texting to humans for approval; an automated job notice for suitably qualified and pre-registered engineers to tender a bid against; applications by many engineers and acceptance of one, include a click-agreement contract offer.
This is followed by contract acceptance and in-part prepayment through a block chain account…
Then - just minutes later - the engineer has arrived onsite and is scanning the area with a UAV comprising a multi-sensor payload collecting UV, visible, infrared, radio, X-ray and acoustic data of the target area; an instant later with the data all processed and the problem confirmed; a price of replacement parts is quoted,and accepted.
Delivery of the part - by drone of course - is made with sufficient time for the engineer to watch an 8-minute online video on how best to fit this type of gasket…
With the part fitted by the engineer; the site is resurveyed with the original UAV and sensor payload and both pre- and post-work data-stacks are provided to the well-head management company for verification and audit… Phew!
So, despite this being a clearly over-elaborate scenario used to prove a point in a concentrated way – I take it that you see this is genuinely where we are heading, right?
JB: This sort of thing could definitely happen in a city. For example:
- A water-main will burst on a busy high street;
- The machine learning algorithm failed to predict that a water main would fail - but because it didn’t predict it this time, it shall learn from this mistake;
- The city autonomous vehicles re-route themselves as a network of sensors has informed them of the flood and the road closure;
- There is an asset register of that infrastructure so both people and automated devices can see what other works have been done on that road before;
- The local electricity is switched off for the duration of the repairs but the local electricity storage – either as distributed batteries or autonomous vehicle batteries - take up the slack for the period of outage.
- Robotic weasels enter the pipes to survey the damage and start repairs – possibly as a short term solution and until such time that the fuller works can be scheduled and performed.
I personally believe that this can happen but for it to begin it needs to be motivated by a specific problem - it is unlikely to emerge without a use case. Further, this whole process will be incremental and these capabilities will build up over time. Installing physical infrastructure sensors everywhere would be a nightmare. Recognising synergies between these technologies is crucial and people will have to explicitly go out of their way to identify these things.
Arguably, these types of high-tech workflows exist in vertical silos – the energy industry is a great example - but getting these workflows to operate seamlessly across vertical industries - for example both: energy,waste and water - is a far bigger challenge.
We will need people whose specific role is to develop, promote and commercialise these efficiencies and these horizontal connections. And these people need skills – skills including technological competency, but crucially with a spread of domain expertise.
Joe Bailey is a data scientist at Future Cities Catapult.
Future Cities Catapult accelerate urban ideas to market, to grow the economy and make cities better. They bring together businesses, universities and city leaders so that they can work with each other to solve the problems that cities face.
As part of the lab at Future Cities Catapult, Joe is currently working on how modelling can be used to make better information-driven decisions.
Joe has a articular interest in visualisation and geospatial communication
Andrew Iwanoczko is an EO Specialist and Account Manager at Harris Geospatial.
Harris Geospatial provides technology, analytics development and consulting services underpinning imagery exploitation: Multispectral, Hyperspectral and Synthetic Aperture Radar (SAR) images acquired from satellite, manned fixed-wing airborne and UAV platforms.
Andy’s keen personal interest in Future Cities and the integrated technologies supporting these spawn from his passion for Earth Observation imagery analytics. Through partnership and collaboration, Andrew helps to transform imagery and geospatial data into a crucial cog for Future City design.
Photo Source: https://unsplash.com/photos/bIx15C7AnNg
Categories: ENVI Blog | Imagery Speaks
Tags: Deep Learning; Future Cities; Imagery
Whether that’s from optical, LiDAR or SAR
sensors mounted on satellite, airborne or unmanned platforms - like my
colleagues and associates in this emerging industry, I share a thoroughly passionate
desire to be supporting application development and dissemination of Earth
Observation derived products.
In my capacity as Vice-Chairman for the
British Association of Remote Sensing Companies over the past few years having
it has been important to me to champion the need for routine, high-frequency,
low-overhead, low-value funding grants to support the sector.
It is therefore great to see a shift
towards genuine and on-going support towards tech start-up and SME development,
in parallel with the traditional high-value commercial contracts provided to
support such initiatives as the Sentinel data archive and delivery within the
This summer, we have seen a real flood of
Some of these, such as the Innovate UK: Open Funding Competition,
are quite general in their scope. This
however provides organisations that may simply have a view to a commercial
product derived from Earth Observations in its most embryonic stages, the
impetus to undertake the market analysis and realise their entrepreneurial
Masters, a funding programme jointly orchestrated
between Satellite Applications Catapult, Innovate UK and the UK Space Agency
challenges applications to explore how the integration of satellite Earth
observation imagery with other sources of data may be innovatively used to
create new or improved products and services that address the key challenges
faced by the growing global population.
More highly specific funding grants do tend
to increase the total value of any contract award, but hand-in-hand also rely
on additional overhead as part of the application process in many cases
necessitating the need for a small consortia to directly partner with companies
in-country, and supporting capacity building outside of the UK.
One example is the Innovate UK Funding
competition: Agri-food innovation in Turkey which aims to provide innovative solutions to
challenges faced in the agri-food sector in Turkey.
Whatever your ambitions, it’s great to be
working with so many SMEs and start-ups – some of which have a rather
light-touch background when it comes to earth observation, spectral science and
SAR processing capability. But the thing they all share is a strong ambition, knowledge
of their market and each can see the big prize just a little closer than the
For those that are keen to take the plunge,
and need the soft hand-hold of an industry giant technology provider and product-enabler
- the door is open, the coffee is brewed and the pastry delivery is in!
Image ©2016, European Space Agency,
Download the Environment & Climate Brochure
As someone who is continuously building capacity, products and diversity of commercial offering through partnering, I am confronted by stark approaches to the partnering process.
Whether it’s the perceived threat of:
- a loss of technical commercial advantage;
- losing revenue against existing headcount / overhead;
- a loss of intellectual property (IP);
- a loss of control over that IP; or
- simply poor timing,
there are often many threats that are identified in a SWOT analysis of a collaboration that one or other party feels is too high to continue.
I am hardly a psychologist, but I can appreciate a fundamental aspect of human behaviour that invests time and trust in those that we stand to benefit most from. The flip side being that from those we do not benefit, we exploit. In a personal relationship, especially in the early days of one, whoever cares the least has the power, right?
So it’s no surprise then really that companies are often cautious about partnering. The mood of either party can shift behind the scenes with little apparent evidence shown to the other. Months or years may pass, and the unwitting party finds the partnership has broken down and there is no contingency plan in place to jump to. Meanwhile the former partner leaps forward and moves from strength to strength.
So why bother partnering at all?
Over the past five to seven years, I have seen a change is business climate, largely I am guessing precipitated by global factors that have led to companies consolidating, and focussing on their core strengths.
The most successful companies I've been lucky enough to interact with are those that know their market well, are focussed on their specialisation, and hold close the external parties that enable them to take strides ahead in developing their commercial offering.
Yes, the big guys like to do everything themselves, the perceived lower risk option of keeping everything in-house. But I've seen first-hand the problem of throwing money after money at internal staff - money that the same organisation would never accept putting to a collaboration partner or contracted party for a similar result.
Success for me is measured by profitability rather than total revenue or any gravitas based on the number of staff. Far better to have any number of parties signed up to an initiative where they all have a vested interest and share a piece of that revenue pie. Each commercial success won't necessarily deliver the same bang… but by simply by joining forces there is a doubling of sales and marketing reach than going it alone. And it reduces the risk of competition.
What about the bad eggs?
Yeah, not every party that has a similar ambition and complementary skills is a perfect collaborator. There are plenty of sharks around, and they are pretty easy to spot. These types come into a collaboration discussion seeking to identify what the other party can do for them, rather than putting forward broad opportunities and what competence they bring to the initiative.
The time wasters are of course frustrating. These guys like talking the talk, but never really take time to do the work that goes along with the collaboration. Weeks turn to months with little progress. You initially feel that it might be related to the internal conflict between current work in hand or feeding the beast where revenues are sliding - but in the end the true picture emerges.
I come into collaboration from a position trust rather than distrust. The confidence tricksters are therefore a little harder to shakedown. It's much more subtle. You can (well I can) only pick these up with a constant eye out for odd behaviour...
- Why didn't they query the NDA, MoU or EULA particularly rigorously? Do they have any intention of keeping to it?
- If agreement clauses say one thing, and after being queried are explained in a very different light - Why then is there resistance to permit modifications to reflect this new understanding?
- Once things are all signed up, and commercially sensitive information is shared, and there is movement on one side but an absence on the other - what's going on?
You will spot them in the end, and yes it might be painful...But so is developing a core piece of internal capability and having key staff disappear to more interesting, lucrative, highly-valued positions with different organisations. The risk bar is set low on this circumstance.
What then, is the preferred way to go about partnering?
As I see it:
- be frank about what's in it for you and what's in it for any potential partner;
- ensure the skills are really complementary, and that in other areas you are not in competition too often;
- be clear on the definition of roles both between organisations and for the individuals within each, including which party retains which bits of IP;
- ensure that both parties are signed up in both practically and philosophically;
- accept that much of the collaboration effort in set up may need to be completed out of hours, including both calls and contract / framework review;
- once the wheels are in motion, ensure that there is sufficient internal staff time to manage the ongoing collaboration within core business hours;
- be mindful of each party's revenue model to ensure that whatever is agreed fits the long term plan;
- be flexible and ready to adapt to new opportunities that are not necessarily equitable between the collaborating parties in every case, but ought to be overall;
- keep a mind's eye on the threat of an exit by the collaborating partner.
Exelis VIS UK
Image courtesy of Airbus Defence and Space
Categories: ENVI Blog | Imagery Speaks, IDL Blog | IDL Data Point
Tags: ENVI, IDL, ENVI Services Engine, Earth Observation, Image Processing, Operational Services
Remote sending holds the key to a vast array of applications – if the human eye can detect such subtlety across the visible spectrum of 400nm to 700nm - then it stands to good reason that using clever pixel maths, we can detect a whole lot more across 400nm to 2500nm, right?
Possibly a hangover from academic days, the majority of commercial geo-spatial effort is managed… well undertaken… in the desktop environment. Earth Observation specialists and the organisations they represent are happy to button-click through in the desktop environment. It's agreed that in the workflow development phase, this is entirely sensible and comes well-recommended.
The inefficiency however is in maintaining a button-click style approach to the operational phase of consultancy contracts, or worse, in the production phase of imagery exploitation and product generation.
Spectral scientists and their paymasters have a choice - maintain the inefficiency and utilise their staff in areas where project costs could be reduced, or automate all or most of their processing and harness some combination of: project-margin improvement; greater processing capability for a broader market offering; and/or securing more work with more compelling pricing.
Whether your feature extraction, classification or change detection requires imagery captured from satellites, aircraft of UAV/UAS platforms, it’s typical to require some bespoke workflow generation: weaving together a sequence of parameterised science-led algorithms into a single routine that focuses in on your targeted output.
From my vista, stepping into automation goes something like this:
Where you want to be able to distribute the processing to non-specialists you might even go the distance and develop a Graphical User Interface (GUI) to run locally against the remote sensing tool-kit. This has the benefit of turning administrative and support staff into pseudo-image processing specialist resource!
So let’s get started! In doing so, choose your remote sensing technology carefully by asking yourself and your software vendor the following questions:
If your choice is in using the ENVI Suite of products, automation is underpinned by the IDL development environment. Exelis provides a number of training courses including both Spectral Analysis in ENVI and Extending ENVI with IDL.
Extending ENVI with IDL is a comprehensive three-day kick-start towards automation of manual workflows comprising the use ENVI desktop algorithms and user-developed parameters. It represents the quickest means to go from desktop to automation. The full course outline is available below and the next such course is available in Bracknell, United Kingdom, during May 2015.
Please feel free to get in touch if you’d like to talk more about what’s involved.
Tags: ENVI, IDL, Image Processing, Training, Automation, Spectral Analysis, Extending ENVI with IDL, batch processing
With spatial data, GIS data-processing and geographic information dissemination all-but ubiquitous in the web and mobile app markets, why is remote sensing lagging behind?
Who then is going to fill that gap in the market and convert the long and proud history of image processing of remotely sensed data into hands-off, application-led, auto-magically processed, meaningful information layers?
Remote sensing is itself, nothing new. The toolkits have been developed to support the needs of academia and defence industries over decades. Those tools are steeped in scientific rigour and are authoritative.
When I speak to various individuals within or on the fringes of the Earth Observation and Image Processing landscape about the ENVI Suite of products, broadly, I am speaking to one of two kinds of people
1) Earth Observation Specialists / Spectral Scientists with a knowledge of remote sensing and how to extract, classify or detect “the thing”; and
2) Change-Managers looking to drive commercial efficiency in their business or to enhance an existing capability.
It is typical that a small number of organisations who recognise the value of remote sensing and image processing, have recruited spectral scientists to fill that need as an internal capability. Contracting is used seldom, initiated at times when that earth observation specialist is stretched technically; has limited time; holds sufficient budget; and a can cite a long-list of high-priority tasks in hand.
Either directly, or through its contracting, it is rare that the spectral scientist undertakes work that has broad appeal - mostly conducting effort against internal business needs, or for funded projects and initiatives with a narrow set of deliverable specified by others.
Where workflow development does have some breadth to its appeal, few organisations are presently prioritising the scaling-up of their operations – most organisations are slow to move workflows out of the desktop environment and into enterprise-scale operational-services that might be consumed by wider B2B audiences through OGC web services and REST, SOAP and WSDL protocols.
Change Managers look to remote sensing to solve challenges from a completely different paradigm. Theirs is the position of improvement, be it more bang for the same buck should an innovative culture pervade. Or, where there is a need to drive down costs, change managers look to remote sensing to reduce total resource outlay against existing business practices.
The principal difference is that Change Managers are looking for answers now, without the need for proof of concept and R&D effort and cost. If remote sensing holds the key to the unlocking of their new business process, then they are thinking “the reason that people aren't already doing it, is that it can’t be done”.
Wrong! The real reason is that there is a valley of death between the spectral scientists,those with the financial clout to make the neat products that meet the various domain needs, and the change managers who need the processing capability all figured out, robust and automated and ready-to-use by someone who has never heard of a Principal Component Analysis.
This community of EO Specialists has not yet scratched the surface on the widest breadth of available and commercially sound applications. We are toddlers, who whilst running… and running steadily, think perhaps we are already World Champion Athletes.
So, the challenge I lay down to the remote sensing community is:
- Pick your vertical;
- Invest in understanding that market need;
- Develop workflows for a community, rather than single contracts;
- Automate these workflows;
- Take advantage of the available technology for maximising the processing capability and dissemination methods;
and in doing all of these things, you fill that current void – and bridge the valley of death.
Exelis, through the availability of the ENVI Suite, is well place to assist in a number of aspects across the stages here – flexibly - as much or as little as needed.
Tags: ENVI, IDL, ENVI Services Engine, Earth Observation, Image Processing, Operational Services, Spectral Analysis, batch processing