The selection of an appropriate algorithm depends upon the image content, and type of information required postsegmentation. Many thresholding algorithms are published in the literature, and selecting an appropriate one can be a difficult task. The second approach, known as the triangle method, determines the largest vertical distance from a line drawn from the background peak to the highest occurring gray-level value. The first approach assumes that the background peak shows a normal distribution, and the threshold is determined as an offset based on the mean and the width of the background peak. In these cases, two approaches are commonly used to determine the threshold. Moreover, most images have a dominant background peak present. However, in most biologic applications, both the foreground object and the background distributions are unknown. When the distributions of the background and the object pixels are known and unimodal, then the threshold value can be determined by applying the Bayes rule. Threshold determination from the image histogram is probably one of the most widely used techniques. Castleman, in Handbook of Image and Video Processing (Second Edition), 2005 4.3.2 Threshold Selection However, no parameters are required from the analyst to implement the transformation, making it easy to apply.įatima A. The contrast of an equalized image is often rather harsh, so equalization is not recommended as a general purpose stretch. 5-19 as the variable spacing of GLs in the enhanced image histogram. The highest gain therefore occurs at DNs with the most pixels. Where the CDF increases rapidly, the contrast gain also increases. Because of the unimodal shape of most image histograms, equalization tends to automatically reduce the contrast in very light or dark areas and to expand the middle DNs toward the low and high ends of the GL scale. Equalization refers to the fact that the histogram of the processed image is approximately uniform in density (number of pixels /GL) ( Gonzalez and Woods, 2002). It is achieved by using the Cumulative Distribution Function (CDF) of the image as the transformation function, after appropriate scaling of the ordinate axis to correspond to output GLs. Histogram equalization is a widely-used nonlinear transformation ( Fig.
More than two linear segments may be used in the transformation for better control over the image contrast. The transformation parameters are selected to move the input minimum and maximum DNs to the extremes of the display GL range and to move the mode of the histogram to the center of the display range (128). This example is a two segment stretch, with the left segment having a higher gain than the right segment. With a piecewise-linear transformation, more control is gained over the image contrast, and the histogram asymmetry can be reduced, thus making better use of the available display range ( Fig. If the image histogram is asymmetric, as it often is, it is impossible to simultaneously control the average display GL and the amount of saturation at the ends of the histogram with a simple linear transformation. SchowengerdtProfessor, in Remote Sensing (Third edition), 2007 Nonlinear stretch
0 kommentar(er)
