1. Space Borne Environmental Intelligence
via SUPER high-precision co-registration
of satellite imagery
www.imaginlabs.com
# of Customers
Day 1 11
Day 2 12
Day 3 5
Day 4 10
Day 5 3
Total 41
Sebastien Francois Jiao James Tamara
Leprince (PhD) Ayoub (MS) Lin (PhD) Hollingsworth (PhD) Knutsen (grad)
2. Day 1
Partners Activities Value Prop. Cust. Relations Cust. Segments
• R&D on multi-angle
imagery, image • Precision
• Imagery providers Service for 2D or • “consulting” style
fusion/registration
(DigitalGlobe) 3D sub-pixel agriculture
• Computing engine
dev/testing/releasing? companies/Forestry
registration of
• USGS/NASA • (web) UI
scientists • Dev/admin computing satellite images • NASA: processing of
infrastructure Mars imagery
• Customizing to • USGS: damage
• University
customer needs
researchers using our assessment/earthqu
free software Resources Channels akes
• Scientists / developers • Oil and Gas: sand
• Hardware (cloud?) • Direct sales or via dune risk and
• High bandwidth pipeline routing
distribution partner
network
• Patents and company • Data exchange via
“know how” network (ftp)
• User • Data exchange via
feedbacks/newsletters/ FedEX for large
blogs volume
Costs Revenue
• Wages, office space • Revenue from processing services (subscription, pay per
• Workstations and servers image with volume discount)
• High-bandwidth internet connection
• Cloud services
• Caltech royalties
3. Day 1
• Technology description: Caltech technology to co-register time-series of
satellite and aerial imagery with accuracy better than 1/10 of the image
pixel size.
• Initial hypothesis: Basically, everyone in remote sensing SHOULD be
interested!
• Customer Segments:
– Oil & Gas, Pipeline monitoring
– Disaster assessment
– Agriculture
– Mars imagery processing
• Experiments:
– Discussions with NASA, USGS, USDA personnel
– Discussions with pipeline consultant
4. Day 2
Partners Activities Value Prop. Cust. Relations Cust. Segments
• R&D on multi-angle
imagery, image Automated and fast • Crop-consulting
• Imagery providers • “consulting” style
fusion/registration companies for crop
(DigitalGlobe) service for multi-
• Computing engine monitoring
dev/testing/releasing? spectral image
• vegetable crops = most
• USGS/NASA • (web) UI registration lucrative
scientists • Dev/admin computing • high-end in-season
infrastructure problems faced by crop crop monitoring
• Customizing to consultants: manual companies are
• University
customer needs registration, data interested, but still need
researchers using our timeliness, higher to find good leads at
free software Resources resolution harder to Channels Cargill+ADM
• Scientists / developers register, in-season •Possible interest from
• Hardware (cloud?) analysis • Direct sales or via DEA, difficult to find
• High bandwidth contacts
distribution partner
network • Defense
• Patents and company Service for 2D or 3D • Data exchange via
• NASA: processing of
“know how” sub-pixel registration network (ftp)
Mars imagery
• User of satellite images • Data exchange via • USGS: damage
feedbacks/newsletters/ FedEX for large assessment/earthquakes
blogs volume • Oil and Gas: sand dune
risk and pipeline routing
Costs Revenue
• Wages, office space • Revenue from processing services (subscription, pay per
• Workstations and servers image with volume discount)
• High-bandwidth internet connection
• Cloud services
• Caltech royalties
5. Day 2
• Not everyone cares!
• Too many market segments. Refocus on agriculture only, drop other
customer segments. Agriculture seems to be the most responsive
customer segment with strongest feedback.
• New hypothesis: Crop monitoring needs better image registration to
extract maps of crop/soil variability over time
• Need to better understand the global remote sensing ecosystem.
• Experiments:
– Contacted crop consultant companies
– Contacted several branches of USDA and academic agriculture
institutions
6. Satellite Remote Sensing Ecosystem
WHERE WE
WANT TO BE!
Oil & Gas, Mining Exxon Mobile, Shell, BP, Fugro, etc
Agriculture Crop consultants, farmers, cooperatives,
Satellite Cargill, Monsanto, GeoSys
image Science Research Universities, Research labs
MOST PROMISING
providers Mapping Google Maps, Apple FEEDBACK!
Therefore, focusing
Insurance RMS, Aon, ABS, on this for now…
40%
Annual revenue
Civil engineering Parsons, McCarthy
60%
Annual revenue
Federal
Military: USAF, USACE, NGA,…
Federal Agencies:
applications
USGS, NASA, NOAA, Caltrans,…
7. Day 3
Partners Activities Value Prop. Cust. Relations Cust. Segments
• R&D on multi-angle
imagery, image
• Imagery providers • “consulting” U.S. Crop
fusion/registration
(DigitalGlobe) • Computing engine Problem: manual style consultants
dev/testing/releasing? registration of International
• USGS/NASA • (web) UI
scientists • Dev/admin computing satellite image Crop consultant
infrastructure before comparison
• Customizing to
• University
researchers using
customer needs USDA
our free software Resources Value Prop.: Channels (No registration
• Scientists / developers - automated issue (for now!))
• Hardware (cloud?) - fast • Web access
• High bandwidth
- very accurate • Data exchange Agri.
network
• Patents and company via network (ftp)
“know how”
service for satellite • Data exchange Universities
• User image registration via FedEX for large (lack of money,
feedbacks/newsletter volume only use Landsat,
s/blogs
which has no
registration issue)
Costs Revenue
• Wages, office space, workstations and servers • Revenue from processing services (subscription, pay per
• Cloud services image with volume discount)
• Caltech royalties
8. Day 3-4
• Narrowed focus of customer segment to Precision Agriculture
• New hypothesis: In-season crop monitoring needs better image
registration to extract maps of crop/soil variability over time
• Customer segments:
– International crop consultants
– Insurance, investment banks, commodity funds
• Experiments:
– Contacted farmer cooperatives about in-season monitoring
– Contacted crop consultants
9. What we learned: Need is much smaller than expected
USGS delivers free Landsat imagery Contacted a major player for in-season
with decent registration, sufficient for crop monitoring (interested in precise
basic crop monitoring and fast registration of images from non-
Landsat)
Other satellite imagers are less
widespread, due to costs Provided trial runs on series of images
they could not register. Confirmed that
High-end crop monitoring requires our results exceeded all commercial
non-Landsat satellites (bad registration) solutions
In-season crop monitoring interest is Strongly interested in subcontracting
HUGE registration processing to us
Potential revenue $75k-$250k/year
Airborne or ground-based systems need more crop consultant companies
compete with satellite imagery for high-
end crop monitoring
USDA delivers crop yield statistics but
does not provides services to farmers
11. Day 4
Partners Activities Value Prop. Cust. Relations Cust. Segments
• R&D on multi-angle
imagery, image
• Imagery providers • “consulting” International
fusion/registration
(DigitalGlobe) • Computing engine Problem: manual style Crop consultant
dev/testing/releasing? registration of for farmers
• USGS/NASA • (web) UI
scientists • Dev/admin computing satellite image
infrastructure before comparison Channels ‘Crop forecast’
• Customizing to
• University
researchers using
customer needs • Web access offices from
our free software Resources Value Prop.: • Fast data Insurance/Invest
• Scientists / developers - automated exchange via ment firms
• Hardware (cloud?) - fast network (ftp)
• High bandwidth • Web access
network - very accurate • Fast data
• Patents and company service for satellite
“know how”
exchange via
• User image registration network (ftp)
feedbacks/newsletter
s/blogs • Data exchange
via FedEX for large USDA
volume Agri. Universities
Costs Revenue
• Wages, office space, workstations and servers • Revenue from processing services (subscription, pay per
• Cloud services image with volume discount)
• Caltech royalties
12. What we learned:
USDA is not a potential client - Insurance companies and
confirmed Investment/hedge funds may contribute
to a significant portion of the revenue
Limitation: too long revisit time of the (WAG)
satellite for in-season. Hardware
limitation. Performed bandwidth experiment
with in-season imagery
Limitation: lag-time between image
analysis results and field decisions. Need Automatic web-based processing is
better integration with the tools of suitable for crop consultants as long as
end-user. ( potential opportunity) data volume does not exceed more than
a few images per day. Otherwise, should
use faster bandwidth to insure timely
Farmers and cooperatives do not seem delivery of information to farmers.
to contribute to the largest revenue from
satellite that needs correction.
13. Day 5
Partners Activities Value Prop. Cust. Relations Cust. Segments
• R&D on multi-angle
imagery, image
• Imagery providers • “consulting” US and
fusion/registration
(DigitalGlobe) • Computing engine Problem: manual style International
dev/testing/releasing? registration of Crop consultants
• USGS/NASA • (web) UI
scientists • Dev/admin computing satellite image and for farmers
infrastructure spectral analyses of Channels
• Customizing to
• University sat im time-series
researchers using
customer needs • Web access ‘Crop forecast’
our free software Resources • Fast data for
• Scientists / developers Value Prop.: exchange via Insurance/Invest
• USDA • Hardware (cloud?) - Fast, automated, network (ftp)
• High bandwidth • Web access
ment firms
network accurate satellite • Fast data
• Patents and company image
“know how”
exchange via
• User registration and network (ftp)
feedbacks/newsletter analysis
s/blogs • Data exchange
- More Intuitive via FedEX for large
Decision Tools volume USDA
Agri. Universities
Costs Revenue
• Wages, office space, workstations and servers • Revenue from processing services (subscription, pay per
• Cloud services image with volume discount)
• Caltech royalties
14. What we learned about Precision Agriculture:
• No universal “black box” solutions • NEED individuals or algorithms who
• NEED better algorithms for: can compile and prepare data from
– Fertilizer rates disparate data sources for Variable
Rate and Site-Specific Ag
– Seeding rates
• Direct benefits of Prec. Ag:
– Herbicide rates
– Increased efficiency in terms of:
• No “cheap” way of getting soil
fertility labor hours,
– Intense manual grid sampling crop yields,
– Direct automatic soil sampling amounts of chemicals and
water needed and applied.
(like AgroBotics AutoProbe)
– Long-term spatial recording of
– indirect sampling technologies operations good for:
(like electrical conductivity) Future land/ag management,
– Remote sensing approaches litigations,
(Airborne, satellite, in-field insurance claims,
solutions) increased farm enterprise value
15. What we learned about farmer needs:
• Farmer interest correlated with kind of crop:
– Lucrative Vegetable/Fruit Crops >> Grains
– Annual Sensitive Crops (Cotton,Tomato) >> Perennials (Vineyards)
• Most farmers WOULD ADOPT satellite crop monitoring if
cheaper and higher temporal sampling (comparisons every 2-
3 days rather than 2 weeks).
16. Day 5
• New hypothesis: In-season crop monitoring needs improvements in image
registration, spectral analyses and algorithms to turn satellite data into
useful decision tools.
• Customer segments:
– International crop consultants
– Insurance, investment banks, commodity funds
• Experiments:
– Contacted crop consultants about in-season monitoring
17. Some numbers… that don’t add-up!
• Imagery costs between $0 (free low res Landsat) and $0.8/km2
• Farmers can afford up to $1-2/acre/year on imagery for pasture
crops (cheapest scenario) with ~2 week sampling equiv
$9.5/im/km2!
• Nevertheless: many dedicated agriculture satellites exist
Who really uses them? (Hunch: Insurance and Investors)
• Need better knowledge (more customer investigations) of who
spends the most on satellite imagery for agriculture
• Essential information to estimate the market potential is missing
18. Next Steps…
• Customer Development Process with insurance and financial
institutions (need more contacts like van Beurden Insurance in
CA)
• Cust. Dev. to explore interest/need in Integrated
Data/Decision Tools and Long-term Spatial Data Management
for Farmers
• Figure out how to get lower cost, more frequent sat images
for ag use.