Terms in this set (38)
Nearest Neighbor Interpolation
Resampling method for geometric correction. Takes value of pixel closest to the input and assigns that value to the output. Quickest way, but not as accurate. May have a jagged appearance.
Resampling method for geometric correction. Determines the value from the weighted average of 9 closest pixels to the input and assigns that value to the output. Smooth but the values are altered which produces blurring/loss of image resolution.
Resampling method for geometric correction that determines value from the weighted average of 16 closest pixels to input and assigns the value to the output. Somewhat sharp and doesn't have a disjointed appearance
Thin Plate Spline
Mathematical transformation during geometric rectification. Collected GCPs are used to compute the result, warping the image.
Used to create land cover classes during supervised classification. Takes min/max or standard deviation and creates a "boxes" of classes. Any values outside of them is not assigned to a class & there is overlap the smaller class is used. Quite a fast method but doesn't separate pixels as well.
Used to create land cover classes during supervised classification. Classifies according to the closest mean value. Every pixel classified but areas with a lot of variance, like urban areas, may be misclassified. Pretty fast method.
Used to create land cover classes during supervised classification. Uses probability to determine classification. Most accurate and CPU intensive. Slower method.
Convert a raster layer to
a polygon layer. Creating polygons from land cover classes
Calculating the hectares of each land cover class
Unsupervised classification. Creating land cover clusters
-Arbitrary mean values created for classes from min. spectral distance (scatter plot) .
-Fast, produces statistically normal data
-Unsupervised classification. Standard way to create land cover clusters
-For every iteration during iterative clustering: pixel=nearest cluster, clusters with less than x # of pixels deleted, mean values recalculated, exit when x% pixels change. Can be slow.
Fuzzy K Means
Unsupervised classification. Creating land cover clusters.
-No exact boundaries.
-choose # of clusters, assign random coefficients for being in cluster, repeat until change no more than x.
Smoothes out noise (speckles) while keeping sharp features/edges.
1. Enhanced Frost (FEFROST)
2. Enhanced Lee Adaptive (FELEE)
3. Gamma (FGAMMA)
4. Kuan (FKUAN)
5. Lee Adaptive (FLE)
RADAR Filter. You can use this filter for lots of ________ distributed images ie. forested areas, agricultural lands, and oceans. Preserves the observed pixel value for non-_______-distributed images. Keeps texture.
Non linear Contrast Stretch (RSTR)
Used on SAR images. Generates a lookup table (LUT) segment to perform a _____________ designed to enhance 8-bit radar imagery.
RADAR Textures (TEX)
calculates a set of measures for all pixels in an input image. Based on:
Texture produced based on areas that have different contrast
Texture produced based on areas that are different
Texture produced based on areas that are the same
Enhanced Frost (FEFROST)
-homogeneous areas: speckles removed using low-pass filter.
-heterogeneous areas: speckles are reduced with convolution kernel.
-areas containing isolated point targets: observed value preserved.
Enhanced Lee Adaptive(FELEE)
-minimizes loss of radiometric and textural information.
-Homogeneous: The pixel value is replaced by the average of the filter window.
-Heterogeneous: The pixel value is replaced by a weighted average.
-Isolated Point target: The pixel value is not changed.
-minimizes the loss of radiometric and textural information.
-transforms the multiplicative noise model into a signal-dependent additive noise model. then min. mean square error criteria is applied to the model.
Lee Adaptive (FLE)
-smooth speckles that have an intensity related to the image scene & an additive and/or multiplicative component.
-standard deviation based, uses statistics to calculate
-pixel being filtered is replaced by a value calculated using the surrounding pixels.
DEM. Creates a vector segment containing _______ lines from a raster image, such as an elevation image, given a specified ________ interval.
DEM. calculates the surface ______ for each pixel of an elevation channel. The output image contains slope values that range from 0 to 90 degrees.
DEM. calculates surface _______ (orientation) angles from an elevation image. Angles represent directions in which the slopes are facing.
Shaded Relief (REL)
DEM. Define light source to convert an elevation channel to a _______ image. Values are scaled from 0 to 255.
typically used to convert data from high-resolution (32- and 16-bit) channels to low-resolution (8- and 16-bit) channels
Performs arithmetic operations between two image channels.
Performs arithmetic operations between an image channel and a scalar constant.
Intensity, Hue, Saturation (IHS)
Used in fusion. Forward color transformations, outputs ____________ images.
Red, Green, Blue (RGB)
Used in fusion. Inverse color transformations, outputs ____________ images
fuses RGB with a B&W intensity image. Outputs RGB with same resolution as the original B/W intensity image, but where the color (hue and saturation) is derived from RGB.
outputs RGB by fusing an input RGB or PCT with B&W intensity, using the IHS transform (Cylinder or Hexcone model) or the Brovey transform.
Fuses PCT image with B&W intensity image. Outputs RGB with the same resolution as the original B/W intensity image, but where the color (hue and saturation) is derived from the PCT
Line Replacement (LRP)
Radiometric Correction. Corrects single line dropout using the average DN value of pixels above and below.
Radiometric Correction. Corrects the cyclical striping effects that were caused by different detector signal responses.
Radiometric Correction. Creates bitmaps from image channels, given minimum and maximum _________ values.