Imagery and Its Use in GIS


Share:

Imagery is a type of data that is useful for many GIS applications and is defined as any type of photograph. Normally photographs used for GIS projects consist of images gathered from a satellite or an aircraft. These can be photos that are printed on film or they can be taken with a digital camera and stored as digital images (NCSU Libraries). In addition, printed topographic maps that are scanned as digital images also constitute imagery.

In terms of the specific GIS data type, imagery is considered raster data. As such, all GIS images are made up of a grid of numbers that are arranged into rows and columns. Each grid is called a pixel and the pixels are assigned different numeric values. The different values assigned to the pixels typically represent quantities that identify things like elevation, slope gradient, or spectral brightness of an area (Verbyla, 1995). When all of the pixels are combined an image is formed.

Imagery is a data type that is extremely useful for GIS. It comes in many different types and it is able to show both large and small areas in varying levels of detail. This makes it versatile for a wide variety of different GIS projects and as such it is a common type of data for GIS projects.

Types of GIS Imagery

Imagery that is used in GIS can include aerial photos, satellite images, thermal images, digital elevation models (DEMs), scanned maps, land classification maps and surfaces created from a previous analysis that are saved as an image (Managing Imagery with ArcGIS 10). Hyperspectral, multispectral and panchromatic are general terms that describe imagery types. Hyperspectral imagery is imagery that is used for classifying different land types on the Earth (Dempsey, 2011). It is mostly used for agriculture, forestry management and other projects that examine the Earth’s physical landscape. Multispectral imagery is imagery that is made up of two or more images that are taken at the same type but in different portions of the electromagnetic spectrum.

Images gathered via satellite are some of the most commonly used images in GIS. There are two main types of scanning systems that obtain the various types of imagery used in GIS today (Verbyla, 1995). These are whiskbroom and pushbroom scanning. Whiskbroom scanning uses a scanning mirror that quickly sweeps back and forth over an area while taking an image. This type of scanning is used in the Landsat  Thematic Mapper and Multispectral scanners to obtain satellite imagery. Pushbroom scanning is currently in use in the SPOT satellites and uses a row of silicon detectors to take images as the satellite flies over an area (Verbyla, 1995). Aerial photographs taken from planes are another method of obtaining imagery for use in GIS.

Click and drag the slider bar to compare these LDCM images, which zoom into the area around Fort Collins, Colo. On the left, the image is shown in natural color, created using data from OLI spectral bands 2 (blue), 3 (green), and 4 (red). The image on the right was created using data from OLI bands 3 (green), 5 (near infrared), and 7 (short wave infrared 2) displayed as blue, green and red, respectively. In the left-hand natural color image, the city's elongated Horsetooth Reservoir, a source of drinking water, lies west of the city. A dark wildfire burn scar from the Galena Fire is visible just to the left of the reservoir. The scar shows up bright, rusty red in the false color image.

LDCM images of the area around Fort Collins, Colorado. L: the image is shown in natural color, created using data from OLI spectral bands 2 (blue), 3 (green), and 4 (red). R: this was created using data from OLI bands 3 (green), 5 (near infrared), and 7 (short wave infrared 2) displayed as blue, green and red, respectively. In the left-hand natural color image, the city’s elongated Horsetooth Reservoir, a source of drinking water, lies west of the city. A dark wildfire burn scar from the Galena Fire is visible just to the left of the reservoir.

Accessing and Storing GIS Imagery

As with any type of GIS data being able to access the correct type imagery data needed is important to conducting a successful analysis. To access the plethora of imagery data that is available for use in GIS one has to understand three different techniques. These include direct access, access from a dynamic service and access from a static service (Visualizing and Analyzing Imagery with ArcGIS 10). Each method of access can provide a multitude of images.

Direct access involves using images that are stored on a personal computer, a local disk or server. Most often images obtained via direct access are stored on the hard drive of a personal computer. When images are shared among multiple users they are sometimes stored on a server or removable storage device that can be accessed on a computer. Access from a dynamic service involves specific data needed from an imagery provider. Images that are accessed from a dynamic service are often created when it is requested and as a result this method is time consuming and expensive. Image access from a static service involves the use of online via databases and web services. Government agencies such as the United States Geological Survey and websites produced by private companies such as ArcGIS.com and Google Maps are static image services. Images from such websites are predefined for specific areas and they are often digitally enhanced and georeferenced for use in GIS projects (Land Trust GIS, 2011).

Once the desired image is accessed via direct, dynamic or static access it is important to correctly store it so that it can be used effectively in a GIS application. The most common ways to store data that consists of many images is in a raster dataset or mosaic dataset. Images stored as a raster are images that are stored individually, while those stored as a mosaic dataset are images that are stored as a collection.

Managing and Using GIS Imagery

When imagery is properly stored this makes it easier for GIS users to begin a variety of different image processing techniques. One of the most common of these is image classification, of which there are two main methods: supervised and unsupervised. Supervised classification is the most common image classification method used in GIS. In this process, the image pixels are categorized by specifying numerical descriptors of the land cover types present in the image and each pixel in the image is then labeled with the name of the category it represents (Verbyla, 1995). The user comes up with the classifications based on known vegetation types.

Unsupervised classification is based on a computer algorithm that groups pixels with similar characteristics into clusters based on user-entered statistical criteria (Verbyla, 1995). Once classified, the user can label the features in the image to show different classes such as vegetation type or a land-use category.

Once an image is classified, it is important for GIS users to then check the accuracy of the image prior to processing and interpreting it. To determine accuracy, a user examines the image to ensure that the information classified in the image matches that of the landscape with other classified documents that are said to be correct. In addition to checking classification accuracy, many GIS users also check precision or the level of measurement for categories in an image. If the precision is very high it is harder to ensure accuracy because some items in the image can be misclassified so it is important to find a happy medium.

After being classified and determined to be accurate images are finally processed and interpreted as GIS projects. Image processing in GIS involves performing operations on raster layers using one or many of the spatial analysis and/or image processing tools available in software packages such as ArcGIS. Hillshading, the creation of a viewshed and NDVI, and pan-sharpening are some commonly used image processing techniques (Chang, 2012). These however, are also just a few of many available in GIS.

NDVI image taken with small unmanned aerial system Stardust II. Mosaic of 299 images acquired in one flight.

NDVI image taken with small unmanned aerial system Stardust II. Mosaic of 299 images acquired in one flight.

Project Examples Using Imagery

Because of the large volume of imagery data available and the tools that exist for processing it in GIS, the possibilities for projects involving imagery data are endless. Imagery data used in GIS can consist black and white USGS aerial photos, highly detailed color satellite images, or USGS land cover data to name just a few.

Recent research projects involving imagery include an examination of the world’s tallest eucalyptus trees, how a tropical cyclone damaged rice production in Myanmar, and an investigation of flooding on the Red River in Manitoba, Canada (ESRI, 2010). These and other projects are significant because they show that imagery is an essential component to GIS.

To see some examples of imagery data that is available for GIS applications visit the Terraserver website.

References

Chang, Kang-tsung. (2012). Introduction to Geographic Information Systems. McGraw-Hill: New York, 6th edition.

Dempsey, Caitlin. (9 March 2011). “Satellite Imagery.” GIS Lounge. Retrieved from: http://www.gislounge.com/satellite-imagery/ (Accessed 28 February 2014).

ESRI. (September 2010). “GIS Best Practices: Imagery.” ESRI. Retrieved from: http://www.esri.com/library/bestpractices/imagery.pdf (Accessed 28 February 2014).

ESRI. (n.d.). Managing Imagery with ArcGIS 10. ESRI Training Seminar Personal Notes. (Taken 30 November 2013).

ESRI. (n.d.). Visualizing and Analyzing Imagery with ArcGIS 10. ESRI Training Seminar Personal Notes. (Taken 30 November 2013).

Land Trust GIS. (2011). “Use Aerial or Satellite Imagery.” Land Trust GIS and GreenInfo Network. Retrieved from: http://landtrustgis.org/technology/advanced/imagery (Accessed 28 February 2014).

North Carolina State University Libraries. (n.d.). “Digital Aerial Imagery and Orthophotographs: GIS: NCSU Libraries.” North Carolina State University. Retrieved from: http://www.lib.ncsu.edu/gis/orthophotos.html (Accessed 28 February 2014).

Verbyla, David L. (1995). Satellite Remote Sensing of Natural Resources. CRC Press, Lewis Publishers: Boca Raton, Florida.



Advertising

Like this article and want more?

Enter your email to receive the weekly GIS Lounge newsletter:


Advertising