Feature

New Application: GIS in the Horizontal Perspective

by Jeff Thurston

Background

Horizontal or side-view GIS is a technique that allows for the development of side-view thematic content suitable for GIS applications. Most if not all GIS applications are developed from a top-down perspective - similar to taking aerial photographs. Horizontal GIS is based upon thematic content originating from the ground. The thematic content is derived parallel to the earth''s surface similar to using a regular hand-camera view perspective. In fact, digital cameras provide one method for acquiring information suitable for use with this technique.

Horizontal GIS offers opportunities for studying environmental, structural change and dynamics such as light and air quality. Often these observations cannot be accurately assessed from top-down traditional perspectives. This may be due to scale, resolution or factors relating to inaccessibility such as under tree canopies or due to clouds. Horizontal GIS may even be applied within buildings. Horizontal GIS is by definition large-scale (< 1:1000). Therefore it is applied in specific areas and is highly associated with individual applications. In operation it can involve location based technologies and other technologies and therefore is also considered an integrated application.

In this column I will discuss some of the concepts of horizontal GIS, how it can be applied, application consideration, its value and issues related to technology integration.

Concept:

Fig.1 - Horizontal View GIS
Tiergarten, Berlin, Germany

Horizontal GIS can be used to monitor changing local environments. The easiest way to imagine horizontal GIS is to think about taking a picture with a digital camera on the landscape - the camera view parallel to the earth (Fig. 1). Indeed, if a series of pictures are taken from the same place over time, a record of these changes are recorded. The temporal and spatial change can then be assessed. Horizontal GIS utilizes these series of pictures

Many environmental models including environmental, agricultural and forest models attempt to simulate local biodiversity and processes using small-scale top-down perspectives. This is due to the scope of these models, which seek to understand and emulate large tracts of land. It is also related to cost. Very large-scale highly accurate data is expensive to collect, maintain and analyze. Accordingly horizontal GIS is considered a means to supplement already existing small-scale analysis.

Fig 2. Horiztonal GIS location and FOV

These areas are usually obstructed which prevents data collection using other methods. Horizontal GIS is also useful for monitoring changes that occur perpendicular to the earth''s surface. These changes can, for example, include lighting, disease, movement and other phenomenon readily apparent using traditional data capture methods. Since the technique is applied perpendicular to the earth''s surface, it is also useful for studying spatial features that are present in the vertical plane. The monitoring of historical buildings and their structural or surface change could also be monitored using side-view or horizontal GIS. The key to developing horizontal thematic layers is in being able to re-locate the same (X, Y) coordinate over time. For short studies involving hours or a day, a tripod with camera can be fixed to a location and pictures taken at regular intervals. For longer periods of time, such as in monitoring environmental changes, the (X, Y) coordinate can be found using GPS (Fig. 2).

The field of view (FOV) or the area that each image represents between visits must be similar. That is, both coordinate and FOV azimuth should be the same. A compass can be used to determine azimuth. In a traditional top-down satellite or aerial photograph the ''nominal'' or average scale is usually used for scale determination. In horizontal GIS applications the scale is variable. Those objects in the foreground are large-scale, while those further toward the horizon are small-scale. This poses unique problems when attempting to measure features from the captured image. How can one tell if two similar trees are the same size when one is closer and the other more distant? Or the row houses along a street that are known to be of similar size, but do not appear so in the image? Thus a need exists to determine scale through the FOV.

New Application: GIS in the Horizontal Perspective Continued…

The scale can be determined at selected interval distances from the camera by using a laser rangefinder. In application a series of features are measured for distance and several ''nominal scales'' can be used if desired. Since the features exist in database they can then have a corresponding distances (scale) assigned to each of them. Upon query only those within a range(s) of scale could then be analyzed.

Fig 3. Horizontal classified (7 classes)

Processing the Image:

An image was taken with a Canon A20 digital camera using 1600 X 1200 resolution. The focal length of the camera is known. A laser range finder is used to determine distances for a series of features at similar distance and noted. The image is classified (7 classes in this example) using ''natural breaks''. A larger or smaller number of classes may be used dependent upon spectral response, resolution and features. The image [TIF] is first imported into a GIS. For this example I used ArcView 3.2 with the Image Analyst Extension - other image analysis products can be used. In this example the idea is to monitor the changes in the undergrowth structurally. Seven classes can be seen (Fig.3). The grid-code values (0 - 255) are classed. The features are then queried by grid-code and laser distance (nominal scales). Each thematic layer is saved for further query and analysis. The independent layers can then be used to generate a TIN or exported as a DXF file or queried (Fig. 4).

Fig 4. Class 3 shapefile in horizontal

Considerations:

Horizontal thematic layers are difficult to classify because they can be affected by both natural and artificial light. The same (X, Y) coordinate and FOV can be readily determined using GPS. The azimuth of the FOV is determined using a standard compass. Since classification will be affected by existing light, that does not preclude classification for images taken one day to the next (assuming clear skies) but may be more problematic over longer periods of time where the sun angle changes. Nevertheless, sun angle can be changed prior to classification from within the software.

This technique primarily addresses structural changes. These include density of trees and composition in a forest or neighborhood features. It may also be useful for the occurrence of phenomenon measured with sensors and instrumentation, such a sound or noise that is being determined in a horizontal fashion. As (Fig. 4) clearly shows, I was able to identify the trees easily and quickly using this method. A closer analysis may include neighborhood pixel relationships and supplemental remote sensing analysis techniques together with a greater number of laser rangefinder measurements. However such detail for a forester or environmental specialist may not be realistic where hundreds of similar measurements have to be made during the course of one week. Normalized Difference Vegetative Index (NDVI) is based upon visible and near-red reflections and can be applied readily with this method. Other soil-water-vegetation indices may be useful for evaluating these types of data. There are other applications and approaches that can be considered with horizontal GIS that are of a integrated nature.

Integration and Applications:

Emergency Applications - let''s assume that there is a release of toxic fumes at a location (X, Y). Nobody knows the nature of the release and sensors are brought to determine (X, Y and Z) the area affected. Where is the source? Is it on the ground or in the air and at which height? Is it evenly distributed or is it unevenly distributed? Horizontal GIS in the sense is then being applied for the purposes of collecting information in a 3-D manner. The value of this is that this information can then be coupled to dispersion models and rates and direction of spread (and assessment) continually monitored. In this example air quality instrumentation is being coupled to GPS, GIS and laser technologies.

Neighborhood Indicators - horizontal GIS can be used to ascertain the numbers and density of foliage on city blocks from equipment operated in a moving vehicle. Overgrown trees will likely obstruct light levels (that be measured) and lower light levels can be associated with higher incidence of street crime. An assessment of street foliage can be used to identify higher risk areas and the information provides an assignment means for engineering crews to tend to the trees. Similarly, assessment of building surfaces may provide indications about the housing in the area.

Fig. 5 SideView of Hoover Dam
Courtesy: Optech Manufacturing

Historical and Commercial Buildings - monitoring historical buildings and their structural surfaces is important for the purpose of preservation and maintenance. Horizontal GIS could be useful for determining surface changes (color or material). This type of application is closely related to heat loss. Through the use of infrared sensor technology heat loss can be ascertained and records provided to identify exposed areas and increase efficiency. For these types of applications, a laser scan of the 3-D structure may be useful where information gathered through horizontal GIS is applied to the ''textures'' - this then becomes a hybrid visualization approach (Fig.5).

Location Based Servicing - These applications could be revolutionary. Since horizontal GIS can be applied by individuals relatively inexpensively these applications assume individuals are contributing to and part of the data capture efforts. Cars already have GPS, what if they had a small digital camera mounted inside? Imagine individual cars gathering useful horizontal GIS information that is then transmitted to a centralized server that processes the information for a region. This information is related to ''travel time'' - numbers of cars, accidents or pedestrians present. Or even the travel speed. Structurally cars, accidents and people would be easy to identify using side-view techniques. Ultimately the quickest route (and secondary) routes could be chosen using horizontal GIS.

Forestry - as indicated in the examples, horizontal GIS could be quite valuable for environmental and forest related studies. Forest inventory information can be ascertained from top-down image platforms. Alternatively, forest sampling is used and ground crews sample selected areas of forest. Horizontal GIS applications may provide a means to sample more areas quicker (higher sampling intensity) as compared to traditional sampling methods.

Music and Arts - music and visual arts are perceived from the side. Lighting, shading and structures can affect that perception. Understanding how sound travels (radiation and reverberation) within local environments can be measured using sound measuring devices. Horizontal GIS can then be applied, locating areas of good versus poor quality sound for all locations within a structure (X, Y and Z). Achieving optimum acoustics in rooms designed for communication is particularly important. Lighting levels can also be determined throughout a structure in the horizontal plane. Using horizontal GIS these levels can be determined and may prove useful for art galleries, enhancing the visual experience or even monitoring climate within buildings. Alternatively, the measurement of light levels in the horizontal plane may be useful for studying (preventing) criminal activity.

Conclusion:

Horizontal GIS is a technique that is useful for side-view applications where temporal and spatial changes are monitored. Through the use of digital imagery coupled to GIS and laser technologies, thematic content is developed. Horizontal GIS is generally applied for very specific studies in a large-scale manner. It can, however, be coupled to top-down traditional GIS capture methods. As such those small-scale traditional applications and analysis methods are used to first identify areas requiring more intensive investigation or measurement. Issues of scale require attention when applying horizontal GIS due to the variable FOV. Applications for horizontal GIS could include the entertainment, engineering, historical, environmental and sociological areas. There is potential for 3-D visualization development using this technique.

Jeff Thurston - is European Director, Integral GIS and based in Berlin, Germany. He holds an MSc. in Geographic Information Systems and advises companies internationally with respect to new opportunities and integrated GIS applications. He is currently writing a book entitled "Integrated GIS / GPS and Geo-technology" to be published by Wiley Publishing, New York in 2003. (jeff@integralgis.com)