4.2 Raster Symbology and Stretch Types (Min-Max, Standard Deviation, Percent Clip, Colormap)
Key Takeaways
- Stretch renderers (like Min-Max or Percent Clip) are used for continuous raster data, enhancing contrast by mapping pixel values across the full 0-255 display range.
- The Classified renderer groups continuous raster data into discrete thematic categories using classification schemes similar to vector Graduated Colors.
- Unique Values renderers are strictly used for discrete thematic rasters, such as land cover or soil types.
- Percent Clip stretches remove extreme outlier pixels (e.g., the top and bottom 1%) to prevent them from skewing the color distribution.
- Standard RGB composites like Natural Color (4-3-2) and Color Infrared (8-4-3) are critical for interpreting multispectral imagery.
Raster Symbology Fundamentals
Visualizing raster data in ArcGIS Pro requires a different approach than vector data. Rasters consist of a grid of cells (pixels), where each cell contains a value representing information such as elevation, temperature, spectral reflectance, or land cover category. Raster symbology hinges on whether the data is continuous (e.g., elevation, imagery) or discrete (e.g., thematic categories like land use).
Raster Renderer Types
ArcGIS Pro provides several renderers specifically designed for raster datasets.
Stretch The Stretch renderer is the most common symbology for continuous, single-band raster data, such as a Digital Elevation Model (DEM) or a single-band grayscale image. It applies a continuous color ramp to the range of pixel values. Because raw raster values often do not occupy the full display range of the monitor (typically 0-255), the Stretch renderer applies a mathematical stretch function to enhance contrast and make subtle variations visible.
Classified The Classified renderer groups continuous single-band raster data into distinct, discrete classes. It operates similarly to the Graduated Colors vector symbology. You choose a classification scheme (e.g., Natural Breaks, Equal Interval) and a number of classes. For example, you might classify a continuous elevation raster into 5 discrete elevation zones, assigning a distinct solid color to each zone.
Unique Values The Unique Values renderer is used for discrete (thematic) raster data, where cell values represent categories rather than continuous measurements. Examples include land cover classifications, soil types, or zoning districts. Each unique pixel value is assigned a specific color. This renderer is only available if the raster has an attribute table (which requires integer values).
RGB Composite The RGB (Red, Green, Blue) Composite renderer is used for multiband raster datasets, such as satellite imagery or aerial photography. It assigns three specific bands from the raster to the Red, Green, and Blue color channels of the display. By altering which bands are mapped to which channels, you can highlight different environmental features.
Vector Field The Vector Field renderer visualizes magnitude and direction data, commonly used for meteorological or oceanographic data, such as wind speed and direction or ocean currents. It uses symbols like arrows, where the size of the arrow represents magnitude and the orientation represents direction.
Discrete The Discrete renderer is similar to the Unique Values renderer but is used when you have continuous data that you want to display with random, discrete colors without formally classifying it. It is rarely used in finalized map products but can be useful for quick data inspection.
Raster Stretch Functions
When using the Stretch or RGB Composite renderers, applying a stretch function is crucial for image enhancement. Most raw imagery looks washed out or overly dark because pixel values cluster in a narrow range. Stretch functions redistribute these values across the full display spectrum (0-255).
| Stretch Type | Description | Best Use Case |
|---|---|---|
| None | No contrast enhancement is applied; original values map directly to display values. | Data already scaled to 0-255, or when strict quantitative display is required without enhancement. |
| Min-Max | Maps the absolute minimum value in the raster to 0 and the maximum to 255, scaling everything linearly in between. | General enhancement when there are no extreme outliers in the data. |
| Percent Clip | Cuts off a specified percentage of the highest and lowest values (e.g., top 1%, bottom 1%) and stretches the remaining values. | Imagery with extreme, anomalous high or low values (outliers) that would otherwise skew a Min-Max stretch. |
| Standard Deviation | Calculates the mean and standard deviation of the pixel values, stretching values within a specified number of standard deviations (usually 2 or 2.5). | Normally distributed data. Highly effective for bringing out detail in typical aerial photography. |
| Histogram Equalization | Distributes values so that there is an equal number of pixels assigned to each display color value. | Enhancing contrast in areas where pixel values are heavily concentrated. Can cause unnatural, exaggerated contrast. |
| Custom | Allows the user to manually define a piecewise linear stretch by manipulating breakpoints on the histogram. | Advanced enhancement requiring precise control over specific value ranges. |
Exam Tip: Understand the distinction between Min-Max and Percent Clip. If a satellite image contains a few extremely bright pixels (glare) and mostly dark pixels, Min-Max will leave the image looking dark. Percent Clip will discard the bright glare pixels from the stretch calculation, vastly improving the contrast of the darker areas.
Gamma Correction
Gamma correction is an additional display adjustment that controls the overall brightness and contrast specifically in the midtones of an image, without significantly affecting the extreme black and white points. A gamma value greater than 1.0 lightens the midtones, while a value less than 1.0 darkens them. Gamma adjustments are often used in conjunction with stretch functions to fine-tune image appearance.
Multispectral Band Combinations
Multispectral imagery (like Landsat or Sentinel) contains data in bands outside the visible spectrum (e.g., Near-Infrared, Shortwave Infrared). By assigning different bands to the RGB display channels, analysts can detect features invisible to the naked eye.
Common combinations (using standard Landsat 8/9 band designations):
- Natural Color (4-3-2): Maps Red to Red, Green to Green, and Blue to Blue. Displays the image as the human eye would see it.
- Color Infrared (8-4-3 or 5-4-3): Maps Near-Infrared to Red, Red to Green, and Green to Blue. Healthy vegetation strongly reflects Near-Infrared and appears bright red, making this combination critical for agricultural and forestry analysis.
- Shortwave Infrared (7-6-4): Excellent for distinguishing moisture content in soil and vegetation, and penetrating thin clouds or smoke.
Choosing the right raster renderer, stretch function, and band combination is essential for extracting meaningful visual information from raster datasets in ArcGIS Pro.
Which stretch function is most appropriate to use if a single-band raster image appears overly dark due to a small number of extremely bright outlier pixels?
Which raster renderer is designed specifically for discrete, thematic categorical data such as a land-use classification raster?
In a standard Color Infrared (CIR) band combination for satellite imagery, healthy vegetation typically appears as which color?