As humans, the colors we see are made up of combinations of reflected wavelengths throughout the visible portion of the electromagnetic spectrum. Each feature that we see has its own unique spectral reflectance curve (i.e. grass, water, cement, etc). These curves are defined by the varying percent of reflectance. The color we see comes from the wavelengths, which are, reflected the most. For example, a green object will reflect high in the green portion of the spectrum, but low in blue and red. In remote sensing, one must understand the reflectance nature of an object if it is going to be identified on an image.Graphs of spectral reflectance curves help us better understand the reflectance nature of an object.
- Acquiring – technology employed (Satellites, aircraft)
- Processing – converting raw data into images
- Interpreting – interpreting or giving meaning to the processed data
Scan photographs or purchase digital data from agencies such as NASA, USGS, BLM, Forest Service or a private company such as SpaceImaging. The data used to come on high density digital tapes (similar to the reel to reel tapes) that had to be read by a tape drive, usually 1 band per tape. Now most data is available on CD's. Small images can be transferred via ftp.
Spatial Resolution
Defining your project and selecting appropriate imagery.
When defining your project, you first decide what you want the imagery for. If it is just for a backdrop, then you could go with aerial photography or panchromatic satellite imagery. If you want multispectral data, then you should be familiar with the different satellites, their spatial resolution and the number of bands each has. The most common are the following:
 |
Note: high resolution refers to rasters with small cell dimensions - high resolution means lots of detail, lots of cells.
Raster and Vector Data
Remember our discussion of the difference between raster and vector GIS?
Raster GIS
Rasterized data divides the entire study area into a regular grid of cells in a specific sequence, the conventional sequence is row by row from the top left corner each cell contains a single value and is space filling.
Cell Values - each pixel or cell is assumed to have only one value. This is often inaccurate because the boundary of two soil types or vegetation and concrete may run across the middle of a pixel. This is called a mixed pixel.
Vector GIS
This model uses discrete line segments or points to identify locations discrete objects (boundaries, streams, cities) are formed by connecting line segments. Vector objects do not necessarily fill space; not all locations in space need to be referenced in the model
Raster GIS
You can either create your own raster data, rasterize vector data, or access digital data in raster format that has already been archived. The latter is the most common way to acquire and use raster data.
When would you rasterize vector data?
To use the data to build a model or perform mathematical calculations. For instance, multiple vector layers could be rasterized to generate values for input into multivariate analysis such as principal components.
However, there are many problems inherent to vector to raster conversions and vice versa, so you must be very careful when deciding to do this. For ex: By forcing real world features into a fixed raster grid, feature boundaries will shift by as much as half the dimension of the grid cells. Small cells create less error but require more storage space.
Or: The typical conversion rule of assigning values using that class which occupies the greatest proportion of the grid cell may result in the deletion of features which are smaller than a grid cell in either the X or Y dimension.
The main strength of raster data is the ability to perform mathematical calculations on the data. There are many mathematical algorithms that can be applied to raster images to pull out the information one is looking for. These can be supervised or unsupervised procedures and are referred to as classifications.
Supervised classifications means that the interpreter selects training site information and the computer algorithm classifies the image based on those sites. Unsupervised classification is where the computer assigns pixels to categories without instructions from the interpreter or operator.
Ex: nearest neighbor. This algorithm looks at adjacent pixels and groups them together based on like pixel values. The operator sets specific parameters as to thresholding and the computer does the rest. Once the classes have been determined, then the operator can go back and make adjustments based on knowledge of the area.
One thing that you may have noticed on the images is that the vegetation comes out in red. That is what we refer to as a False Color Composite, where we have combined bands 4 (near IR), 3(red) and 2(green). Healthy vegetation has a high reflectance in the near (photographic) IR region.
Using GIS and Remote Sensing Data Interactively
As we have mentioned before, the true strength of GIS is in its ability to perform overlay operations between map layers. In cases where map features represent discrete categories, overlay operations can determine the intersection or union of features from different map sources. Maps representing numerical values may also be combined using mathematical relationships. As an example, a GIS may be used to find a good site for a power plant by recoding map layers for soils, slope, and proximity to cooling water and markets into suitability scores or cost estimates. These suitability maps could then be combined mathematically to create a derived map indicating the relative costs and suitabilities for building a facility throughout an entire region.
How could you use remote sensing in Civil Engineering?
1. Erosion prediction – streambank and shoreline erosion management
2. Transportation modeling
3. Dam and reservoir location and planning
4. Geomorphology – channel and watershed characterization
5. Characterizing soil dynamics
6. Geotechnical and environmental engineers
7. Earthquake engineering
8. Construction planning and siting
9. Agricultural engineering
ASSIGNMENT (MS Word format)
Data for assignment
Remote Sensing Powerpoint
