Bentley Descartes has many functionalities to enhance the global appearance of an image, principally by the use of image filters. These filters are found in the Enhance Image tool box which contains the following tools: Histogram, Contrast Stretch, Density Slicing, Blur, Sharpen, Smooth, Detail, Edge Enhance, Find Edges, and Color Table.
The Contrast Stretch tool can be used to increase the contrast of an image.
The goal of the contrast stretch algorithm is to redistribute the range of values taken by the original image to a new, larger range. For example, in the graphic below, the values of the overexposed image are compressed towards the center of the available range. In this case Contrast Stretch can be used redistribute these values to produce an image with proper tone.
The Preview pane allows you to have an immediate visual feedback when modifying a parameter determining the contrast stretch to apply. At the bottom of the preview pane there is a set of command that gives you the ability to explore the preview. From left to right, the commands are : Update View, Restore View, Zoom In, Zoom Out, Fit, Pan Preview, and Update Preview.
The Update View, Restore View and Update Preview commands can be used to control the application of the contrast stretch. Clicking the Update View command will apply the contrast stretch directly in the current view for the selected image. Note that the image on disk is not modified when applying the Update View command. Clicking the Restore View will command will remove the contrast stretch applied when clicking the Update View command. The Update Preview command do essentially the same thing is the first two commands, but for the image in the Preview pane.
The histogram pane (entitled Luminance or Channel depending on the selected edit mode) is placed at the middle of the Contrast Stretch dialog and allows you to control manually the contrast stretch algorithm.
It is possible to control the contrast stretch of an image manually by specifying the shadows and hilites values. The shadows value determines which pixel's value in the original image will be mapped to the lowest possible pixel's value in the enhanced image (i.e. : resulting image after the contrast stretch has been applied). The hilites value has the same function as the shadows value but for the mapping to the highest possible value.
For example, in the graphic below:
The results of applying this contrast stretch is a balanced distribution of pixel values from indicie 0 which is the blackest black to 255 which is the whitest white. This produces an image with a balence tone from the darkest shadows to the brightest highlights.
Productivity Hint : A good result can be obtained easily by moving the black triangle marker at the bottom of the histogram pane to left of the histogram bell shape and the white triangle marker to the right of the histogram bell shape.
The automatic contrast stretch is applied when clicking on the Auto button. The automatic contrast stretch is based on the percentage of pixels with the minimum values or maximum values that should be mapped to the lowest value or to the highest value respectively. Those percentages can be set in the Descartes Settings dialog which is display when clicking on the Settings... button.
Productivity Hint : Use first the automatic contrast stretch for a core adjustment of the contrast stretch to apply. Then use the manual contrast stretch for tweaking the contrast stretch to the desired result.
Density Slicing adds color to an image to increase the visual interpret ability of the map features. This global image enhancement highlights a range of luminance values and assigns it a different color for clarity or for further analysis. This technique canalso be used to slice an image into a series of classifications based on variations of luminance values. For example, the areas of the image that fall between 0-117 may be hydrography, those that are between 118-145 may be prone to flooding, and pixels with values from 146-242 may be areas that have low probability of flooding.
The Blur filter is used to to blur an image. The blurring of an image can be used to remove details or noises and its analog to seeing an image through an out-of-focus lens.
The alorythm Descartes uses is a 5 X 5 convolution matrix in the following form:
[ 1 1 1 1 1 ] [ 1 0 0 0 1 ][ 1 0 I 0 1 ][ 1 0 0 0 1 ][ 1 1 1 1 1 ]
The matrix middle value (i.e. element 2,2) is computed using the following equation :
I = 31 - pi_Intensity * 31 / 255
So increasing the intensity will decrease the weight of the center pixel so as to give the neighbouring pixel greater importance.
The Sharpen filter is used to sharpen the image or, in other words, increase the differences of neighbor pixels' value.
The algorythm Descartes uses is a 3 X 3 convolution matrix om the following form:
[ -2 -2 -2 ] [ -2 I -2 ][ -2 -2 -2 ]
The matrix middle value (i.e. element 1,1) is computed using the following equation :
I = 31 - pi_Intensity * 31 / 255 + 17
Productivity Tip: When excessive sharpening is applied to an image, it may introduce noise or other undesired artifact in areas of graduated colour.
The Smooth tool is used to smooth the image and making it less sharp. The smoothing operation is often used to remove high frequency noises from an image.
Descartes is uses a 3 X 3 convolution matrix in the following form:
[ 1 1 1 ] [ 1 5 1 ][ 1 1 1 ]
Productivity Tip: You can increase the smoothing of the image by applying multiple times the smoothing operation on the same image.
The Detail function is used to increase the details of an image so that they can be visualized and interpreted more easily.
The detail function, while being very simple to use, can give very good result. For example in the left image below, it is very hard to determine the characters on the hot air balloon. To increase the visibility of the characters the detailing filtering was applied three times. The characters are much more visible after the filter has been applied as shown by the right image below.
Original image
Detailed image
Descartes is using a 3 X 3 convolution matrix. The matrix has the following form :
[ 0 -1 0 ] [ -1 10 -1 ][ 0 -1 0 ]
Productivity Tip: You can increase the detailing of the image by applying multiple times the detailing operation on the same image.
The Edge Enhance filter increases the edges of an image by using a high pass filtering (i.e. : filtering out the low frequency information of the image).
The Find Edge functionality is also doing a high pass filtering as the edge enhancement functionality but also darkened the pixels presents in homogeneous part (i.e. : patch of pixels with the same color) of the image. This allows the user to easily visualize the edges in the image.
Descartes is using a 3 X 3 convolution matrix. The matrix for edge enhance functionality has the following form :
[ -1 -1 -1 ] [ -1 10 -1 ][ -1 -1 -1 ]
The matrix for edge finder functionality has the following form :
[ -1 -1 -1 ] [ -1 8 -1 ][ -1 -1 -1 ]
Productivity Tip : While as with the other raster enhancing functionalities it is possible to apply both edge related functionalities more than once it is recommended to do it only once. Doing it multiple times will just increase the small differences of pixels' value in homogeneous area, which is probably not what is wanted in the first place.
The Color Table tool allows you to modify any color of the selected image's color palette.