This is the first technical report for remote sensing of the environment. The report contains information about the learning objectives and imaging functions being introduced and examined in this weeks laboratory study. The lab students completed contained seven sections, which are designed to give a basic introduction to the program ERDAS and several of the image functions. Compiling information on previous labs, lab exercise 4 allowed students to practice delineating a study area, analyzing images by optimizing spatial resolution, along with some radiometric enhancement techniques. Furthermore, students gained experience with various methods of resampling, along with binary change detection using graphical modeling. On top of the many skills listed above, ERDAS has unveiled a fantastic tool to be used as a native platform in ERDAS in which Google Earth is linked to the image being analyzed, providing an unparalleled image analysis suite in a single program.
Methods
The first section of the lab has students designating a specific area of interest (AOI). In the ERDAS program, there is several ways to do this. Firstly, once the base image is loaded into the viewer, the user can utilize the Inquiry box and create the AOI from this box placed on the base image. Although this technique is quite simple, it can only be used if the AOI is contained in a square/rectangle. Often times in dynamic studies, the AOI is rarely a square/rectangle. If an appropriate AOI is some other polygon, there is another path to making a delineated study area. A shapefile (.shp) can be created separately, and imported into the program. Once the (.shp) of the AOI is in ERDAS, the Raster tool "Subset & Chip" can be used to create a delineated study area of the base map, and saved as the specified AOI of the study.
Section three offers a simple radiometric enhancement technique and tool which can be used to reduce the amount of haze in a given image. Haze, or atmospheric scattering gives the image a slightly opaque looks, and essentially degrades the overall information content the image is able to provide. Using Haze reduction enhances the information content of the image while providing the user with a clearer view of the image.
Section four concerns itself with introducing students to a new service provided by Google Earth. Recently, ERDAS provided a native platform for Google Earth to function within the ERDAS programming. This service has many unique opportunities previously unavailable. The main focus of this section briefly introduced methods to sync the view of the image in ERDAS, to the online format of Google earth, essentially linking the image to the precise location of the image geographically on google. This new innovation is a simple tool which provides a great service for projects carried out in ERDAS instead of having to manually locate the image location on google earth.
Section five: Resampling techniques. This section involved the process of changing the size of pixels. Resampling can reduce or increase (not really used), the pixel size to aid in analytic processes. In this lab, students were introduced to two different ways to resample; By the "nearest neighbor" and "bilinear interpolation".
The first technique resamples using information about surrounding pixels, and adopts the values of the surrounding pixels to enhance the resolution of the image. The nearest neighbor technique proves to be a rudimentary technique, resulting in very minimal resolution enhancement. If a more serious effort is needed, the user should implement the Bilinear interpolation which averages neighboring pixels to produce a smooth output which is more spatially accurate. A disadvantage to the interpolation is that the original data is altered, and the overall contrast of the image is degraded. These side-effects are minimal though, and Bilinear Interpolation proves a good method to resampling.
The first technique resamples using information about surrounding pixels, and adopts the values of the surrounding pixels to enhance the resolution of the image. The nearest neighbor technique proves to be a rudimentary technique, resulting in very minimal resolution enhancement. If a more serious effort is needed, the user should implement the Bilinear interpolation which averages neighboring pixels to produce a smooth output which is more spatially accurate. A disadvantage to the interpolation is that the original data is altered, and the overall contrast of the image is degraded. These side-effects are minimal though, and Bilinear Interpolation proves a good method to resampling.
Section six:
