Definition:A radiometric enhancement technique used to improve the visual contrast of an image. It matches the data's dynamic range to the dynamic range of the display medium (photographic or digital).
Explanation:Remote sensing sensor systems detect reflected or emitted radiation from features on the Earth’s surface. Often times, the detected energy does not fill the entire grey level range the sensor is capable of. For example, an image containing water, forest and agriculture may have a range of 183 different grey levels. However, the sensor may actually be capable of recording 256 different grey levels (8 bit system). Some sensors can be sensitive to tens of thousands, millions or billions of grey levels. A contrast enhancement can redistribute these values to utilize the full range of a display. This allows for a greater contrast between features and can lead to a more accurate interpretation of features within the image. There are both non-linear and linear contrast enhancement techniques. Image enhancement techniques are normally used only to improve visual interpretation of an image and its features as the human eye can more readily distinguish high contrast features than low contrast features. Graphic 1 shows an image prior to enhancement (left) and an image after a contrast enhancement has been applied (right). Quantitative processing techniques, such as image classification, change detection or extraction of biophysical parameters from radar backscatter do not benefit from prior contrast enhancing. This is because the computer (as opposed to the human eye) can discriminate between two integer values regardless of how close their values are to each other. A contrast enhancement can be demonstrated by means of the histogram representation of an image. Graphic 2C shows a histogram of raw image data. Graphic 2D shows a histogram for the same image after a contrast enhancement has been applied. Note that in the histogram of the enhanced image, the full range of grey levels is populated. A contrast enhancement can be applied to a selected (grey level) portion of the image. Instead of using the minimum and maximum grey level values in an image, an analyst would specify alternate minimum and maximum values. This approach can be used to enhance the detail of certain features in an image. For example, if there was an urban area located in the image that was the subject of study, then it would be desirable to show as much detail within that region as possible. By applying a contrast enhancement within a specified input range, such as (say) a range of 68 to 217, only the values within that range would be stretched over the entire display range (i.e. 0 to 255). This leads to more contrast within the selected grey range (urban area in this case) and exaggerates minute changes in grey levels. Other features which had grey levels outside this selected range would be displayed as pure black or white. This sort of variation of a contrast enhancement is useful for more detailed study of specific features and is sometimes referred to as a ‘piecemeal’ approach.
Graphic 1:
A - Original Image B - Enhanced Image
Graphic 2:
C - Image Values D - Display Values X - Grey levels Y - Image area (pixels)
Thanks for: http://www.ccrs.nrcan.gc.ca/glossary/index_e.php?id=2889
No comments:
Post a Comment