A histogram (specifically an image histogram) is a graph where the x axis represents the available range of possible values a pixel could have and the y axis represents the total number of pixels which have each specific value. In the example below, most of the pixels have a value in the lower to middle portion of the range.
GeoTIFF images produced by ATLAS are unsigned 16-bit, meaning that in a single band, each pixel could have one of 65,536 possible values.
Usually the pixel values represent the intensity (or brightness) of each pixel. The value "0" represents no light, or the darkest black pixels of the image. The value 65,535 represents the highest intensity of light, or the brightest white pixels of the image. As you can see in the example histogram, there are not many pixels that have the brightest possible intensity (in fact, none of them do).
ATLAS Histograms for Indexes (like NDVI)
ATLAS also allows you to view and analyze your data using various indexes, of which NDVI is the most widely known. NDVI values are not represented by integers and do not represent overall intensity; the values are fractional and range from 0 to 1 (for example, a pixel could have a value of 0.5).
Adjusting the dynamic range can help bring out details of the image, showcase certain aspects of the data, and can even make visible what was once invisible. The histogram controls in ATLAS allow you to adjust or "stretch" the dynamic range by setting custom minimum and maximum values using a slider. The new maximum value becomes the brightest display value and the new minimum becomes the darkest. Pixels above the maximum value will all be displayed with the same intensity as the maximum value; the inverse is true for the minimum value.
You may have noticed that ATLAS uses a color palette (or gradient) to display on the histogram as well as in the map viewer. This helps to showcase subtle differences in values that we would not be able to see in grayscale.
As the histogram range changes, the color palette changes to match so that the new range uses all of the colors of the chosen color palette. As you can see in the example above, all of the pixels above the new maximum (0.15) are colored purple.
This effect can be useful to bring out variation in an area of interest, and could even be used for classification (for example, the various peaks in the histogram often distinguish between soil, road, vegetation).
Focusing on peak from 0.76 to 0.89 (vegetation)
Focusing on the two peaks from 0.01 and 0.15 (road)
Min & Max
You can also enable transparency for values beyond the minimum and maximum range.
The "min" checkbox is deselected, enabling transparency for all pixels smaller than 0.73.