An elevation model is a digital representation of the earth’s terrain. A digital elevation model (DEM) is an important component of many geospatial projects, particularly for engineering, energy and academic studies.
A DEM is created with stereoscopic data, similar to how a 3D movie is made. Let us help you source the raw data necessary for an elevation model that satisfies both your monetary and geospatial project constraints.
As a digital terrain model (DTM) whereby the surface features in a DSM have been removed so that it represents bare earth.
As a digital surface model (DSM) which includes surface features such as vegetation and human-made structures.
Depending on the desired outcomes of your project, the earth’s surface can be represented in two manners:
As a raster grid where each pixel has a known footprint on the ground (e.g., 1 m by 1 m) with an associated elevation (e.g., 525 m). This delivery format is far more common.
As a vector file with continuous lines, or contours, at known increments (e.g., spaced apart every 5 m) with an associated elevation (e.g., 110 m).
USEFUL INFORMATION ON TERMS USED HERE
Circular error 90% (CE90) is the distance within which 90% of the points in the elevation model fall from their actual location in a horizontal or XY direction. Linear error 90% (LE90) is measured the same as is CE90 only it assesses vertical (Z) accuracy.
Ground control points are locations that are visible in the imagery used to create your elevation model with a known latitude, longitude and elevation. They should have an accuracy equal to the pixel size of the input dataset or better. We prefer to use points with an accuracy that is at least half the pixel size.
We suggest extracting contours with a contour interval equal to a 4:1 or 5:1 ratio of the vertical accuracy. For example, a DEM with a vertical accuracy of 1 m LE90 should have contours with a 4 or 5 meter interval. An interval smaller than this represents over interpolated data. Another reliable estimate of the maximum contour interval possible is the relative vertical accuracy of the elevation model.
While there are a variety of factors that control the final quality of a digital elevation model, here are several key factors: (1) the slope of the terrain – i.e. higher slopes can lower the accuracy; (2) the resolution of the raw stereo data – i.e. higher resolution can increase the accuracy; (3) ground control and breaklines – i.e. use of them improves accuracy; and (4) base to height ratio of the stereo pair – i.e. the suggested range is 0.25 to 0.8.
Start For IT can assist you with all aspects of spectral analysis – from the intermediate stages, whereby you can use the processed imagery in your own geospatial investigation; to a final solution that is tailored to address your specific project needs.
COMMON
SPECTRAL ANALYSIS TECHNIQUES
In
the animation above, a red-edge Normalized Difference Vegetation Index (NDVI) has been calculated using pansharpened WorldView-2 8-band satellite imagery.
Red-edge NDVI has been shown to be more consistent and accurate in its assessment of plant health. You will see in the animation that areas of less healthy vegetation have been revealed even though these plants appear to be as ‘green’ as their neighbors. A red-edge NDVI is an invaluable contribution to
any high precision agriculture project.
If your goal is to complete your own spectral analysis but need some help along the way, then our pre-processed products are the correct choice for you.
A Start For IT representative can work with you to determine the final solution that fits the needs of your project. In the meantime, here is a list of the common solutions we can assist you with:
For projects requiring multiple images to completely cover your study site, a mosaic is the ideal solution. Data can be fused together from multiple dates, imagery sensors and even resolutions to create a single, color balanced file that is easy-to-use in ArcGIS, AutoCAD or any other geospatial application.