Ortho Imagery: Pay or Play Free?

June 5, 2013  - By

By Art Kalinksi

esri digitalglobeSeveral weeks ago I attended the ESRI Federal Users conference where Jack Dangermond announced his agreement with Digital Globe to supply the company’s extensive image library to ESRI users. Under the agreement, Digital Globe, which recently merged with GeoEye, is providing its historic imagery library along with very current imagery that is near real-time. The coverage map Jack showed and seen here was quite extensive and covered significant areas around the globe. This prompted a discussion with a long-time GIS colleague, Nancy von Meyer, who suggested that it sure would be nice to have a column that sorts out the current major imagery providers with the pros and cons of the various products and services. So here it is.

In the mid-eighties, when I retired from the Navy and started my second career in GIS, very few users had access to geo-referenced imagery other than intelligence agencies. By the mid-nineties, that changed as digital aerial imagery became cost effective and usable for heads-up digitizing and GIS base map creation. Now there are many imagery collection firms with a lot of content being provided seemingly for free. Too often I’ve heard GISPs fighting budget battles with push-back from non-GIS staffers that goes something like this:  “Why do we need to pay for imagery when we can get Google for free?” Some of the following information and samples may be useful if you need to explain why.

The ASPRS (American Society for Photogrammetry and Remote Sensing) Ten Year Forecast and Survey is a superb source of imagery statistics and trends. Charles Mondello, the project team chair and ASPRS Fellow, highlighted several key points in the survey. By an overwhelming margin, respondents listed ortho imagery as the single most important layer in their GIS work. They also cited the following factors as most critical in making the imagery useful: resolution, positional accuracy, coverage and currency.

Google and Bing

Most ortho imagery is captured by satellites or aircraft and is the dominant type of imagery in use today, with many free sites such a Google or Microsoft Bing offering this imagery. Both websites understand that the better the users’ experience, the more traffic the sites will generate, so they are motivated to constantly improve and update the information. Among the new features being announced by Google is collation of imagery from Google Earth, Google’s Street View, and special projects including its space and underwater imaging. Instead of having to bounce around between products, you’ll now be able to get all of that in one place. Google provides significant support for non-profits and NGOs to map activities worldwide, which builds use and Google’s reputation.

However, the old adage “You get what you pay for” holds true with imagery. If you just want to look at stuff, both Bing and Google are fine, but we must all understand the limitations. By their own admission and licensing agreements, neither Bing nor Google claim to be authoritative GIS data sources. Both are extremely good, but designed for entertainment and to attract users to their sites to expose them to advertising.

A common complaint from Bing and Google users is that they are composites of many different data sources, and they have neither the time nor the need to organize and publish metadata regarding positional accuracy. Metadata is available but very difficult to access.  Users cite positional accuracy estimates of Bing and Google of RMSE of +/- 10 meters. Another issue for federal users is FARS and licensing restrictions. So make sure your legal staff reads the fine print.

Certified ortho rectified aerial imagery

On the other end of the quality and cost spectrum is certified ortho rectified aerial imagery. Well-known companies such as Sanborn, Woolpert, Dewberry, MJ Harden, and a host of smaller aerial imaging firms provide top-notch imagery by contract for specific areas.  Ortho rectified aerial imagery is scale-accurate images that have been corrected horizontally and vertically to account for variations in the terrain. Combining digital aerial images with digital terrain models can cause distortions of elevated features such as buildings, bridges and overpasses. These distortions are corrected during the ortho rectification process. Here is an example from MJ Harden Associates, Inc. showing pre- and post-correction imagery. This is a labor intensive process and not cheap when combined with surveyed ground control, but the end result is a very accurate image on which GIS users can do heads-up digitizing to produce accurate base maps.

mj harden

ESRI and DigitalGlobe

In the middle are imagery vendors that capture imagery for resale and distribution such as DigitalGlobe and others. Some of the imagery is the best currently available, with excellent resolution and positional accuracy and easily accessible and complete metadata. Using the Identify tool in ArcMap or ArcGIS Online, users can see the resolution, collection date, and source of the imagery at the location clicked.  Here is a short matrix of imagery sources and characteristics.

Ortho-Table

Authoritative data vs. Visualization products

It’s very important that we GISPs are mindful of the critical differences between “products” such as DigitalGlobe imagery purchased or delivered via ArcGIS, Bing and Google, and “services” of licensed engineers and photogrammetrists. Many imagery products are superb visualization tools, but should not be confused with authoritative datasets delivered through a licensed and certified process. There is an excellent comparison chart published by the ASPRS that compares the difference between products, professional services and technical services. If your application is critical and has the potential of legal liability, you may want to have your legal department review and understand the differences. It could save you from the false economy of getting cheap or free imagery.

Some examples

The following are some examples of old and new imagery since 1999. This is a non-scientific sample using a piece of property that I bought years ago near Lake Guntersville, Alabama. I’m using it because I know exactly when improvements were made so I can judge the accuracy of the date stamps shown with the images.

Here is the property shown on county-purchased 1996 panchromatic one-meter imagery.

gville 1996

Here is the same property with county purchased 2001 color 2’ imagery.

gville 2001  

This Microsoft Bing Image was taken in early 2011 after a cabin was built but before a garden shed and driveway ramp were added. Looks like 1’ imagery supplied by DigitalGlobe but tagged in Bing as 2013.

gville bing

Google has the most recent imagery, tagged as 2013 but most likely taken in early 2012. This DigitalGlobe imagery looks the best. I’m guessing 6” imagery, but the address marker is off by about 300’.  Not terrible in this area, but could be a real problem with a row of houses.

Blow-out-closeup

So which imagery is best? 

That depends on how you plan on using it. Bing and Google have become so good and ubiquitous that it’s easy to forget that they make no warranties regarding the data quality or suitability for your use. If you are showing friends the location of your favorite fishing spot or displaying a route to a restaurant, Google or Bing will be just fine and most likely better than a boring certified ortho-rectified image. But if the application is critical and puts your organization in legal crosshairs, you may want to use authoritative data produced through a licensed and certified process that has documented metadata.

A good real-world example that I observed several years ago was a failed drug raid in a county near Atlanta. The drug dealers got wind of a potential raid and cleverly switched numbers on their mailbox with a neighbor. The SWAT team did their initial planning on Google and raided the wrong house based only on the mailbox number. Although no one was hurt or killed, the resultant judgment cost the county more than $1 million. That police department no longer relies on free online maps and house numbers. They verify the information using authoritative county 911 data, parcel maps and imagery along with street-level surveillance photos. This was a very costly mistake that might have been prevented. Hopefully your county or agency has heard about this raid and has learned from that mistake.

With the significant expansion of oblique and 3D players, next month I’ll review this growing visualization technology.


UPCOMING GIS WEBINAR

GPS World’s next webinar, “Nightmare on GIS Street: GNSS Accuracy, Datums and Geospatial Data,” is accepting registrations. The webinar will be held Thursday, June 20, 10 a.m. PDT / 1 p.m. EDT / 6 p.m. GMT. Registration is free.

“‘Nightmare on GIS Street: GNSS Accuracy, Datums and Geospatial Data’” is a look at the challenge of dealing with horizontal datums in your GIS,” explained moderator Eric Gakstatter, survey editor and editor of Geospatial Solutions. “We are moving into a new era in dealing with datum transformations. Geodata 2.0 is coming, and it can create big headaches when attempting to combine disparate geospatial databases. Sensors such as GPS receivers, remote sensing imagery, and 3D scanning provide much more accurate data, setting up a collision with outdated and mismatched legacy horizontal datums.”

To learn more about our webinars, visit our webinar page.

This is posted in GeoIntelligence Insider, Opinions

1 Comment on "Ortho Imagery: Pay or Play Free?"

Trackback | Comments RSS Feed

  1. Craig Roberts says:

    Art,

    Thanks for your article, but I was confused by the table. You show a custom certified orthophoto can have a resolution of up to 2″ (5cm), yet its positional accuracy is 2′ (60cm) at best!! I’m no photogrammetrist, but surely this can’t be correct can it? Surveyors provide photocontrol using GPS techniques to 5cm or better. Is this a typo? Please advise.

    Craig Roberts