Identifying groups of pixels that have similar spectral characteristics and determining the various features or land cover classes represented by these groups is an important part of image analysis. This form of analysis is known as classification. Visual classification relies on the analyst's ability to use visual elements (tone, contrast, shape, etc) to classify an image. Digital image classification is based on the spectral information used to create the image and classifies each individual pixel based on its spectral characteristics. The result of a classification is that all pixels in an image are assigned to particular classes or themes (e.g. water, coniferous forest, deciduous forest, corn, wheat, etc.), resulting in a classified image that is essentially a thematic map of the original image. The theme of the classification is selectable, thus a classification can be performed to observe land use patterns, geology, vegetation types, or rainfall.
In classifying an image we must distinguish between spectral classes and information classes. Spectral classes are groups of pixels that have nearly uniform spectral characteristics. Information classes are various themes or groups we are attempting to identify in an image. Information classes may include such classes as deciduous and coniferous forests, various crop types, or inland bodies of water. The objective of image classification is to match the spectral classes in the data to the information classes of interest.
Although any type of image may be classified, multispectral imagery often yields the best results when classified. Classifying a single band is usually very difficult since more than one surface type will have the same digital number. Thus, any spectral class in a single band classification will possibly contain several information classes, and distinguishing between them would be difficult. For classification, two or more bands are normally used, and their combined digital numbers are used to identify the spectral signatures of the spectral classes present in the image. The more bands used to create a classification, the more likely it is that we may get a set of unique land cover classes.
For performing a supervised classification, some prior or acquired knowledge of the classes in a scene is used to identify representative samples of different surface cover types. These samples, known as training regions or sites, are set up to identify the spectral characteristics of each class of interest. The determination of training sites is based on the user’s knowledge of the geographical region and the surface cover types present in the image. Once the training sites have been established, the numerical information in all spectral bands are used to define the spectral "signature" of each class. Once the computer has determined the signatures for each class, it compares every pixel to the signatures and labels it as the class that it is mathematically closest to. Thus, in a supervised classification, the user starts with information classes and uses these to define spectral classes. Each pixel in the image is then assigned to the class which it resembles most closely.
Fig 1 shows a small subset of ASTER data in 14 bands – 3 in the VNIR region, 6 in the SWIR region and 5 in the TIR region, ranging in wavelength from 0.52 μm to 11.65 μm. Fig 2 shows the results of supervised classification of this dataset. The training sites were vectorized, statistics for the training regions calculated and data were classified using the maximum likelihood classifier. All processing was done using ERMapper 7.1.
This website is hosted by
Department of Geology
Aligarh Muslim University, Aligarh - 202 002 (India)