Chesapeake Times, Vol 16 | January 2024
In the early days of SonarWiz, side scan backscatter was converted to 8-bit pixel values on import. The actual intensity values were lost and only an 8-bit color map of the data remained. As sonar systems and sonar users become more sophisticated, SonarWiz switched to importing raw intensity values and today there is a direct mapping between intensity values and colors.
In most respects, the modern coloring system is for the better. Along the way, we jettisoned the Color and Contrast controls of older versions of SonarWiz (See Figure 1) because the values displayed in modern sonarwiz are not image pixels (per say) but actual intensity values.
I was reminded of this old version of SonarWiz while working on two new coloring options we are adding to SonarWiz 8. Both of the features manipulate the one-to-one mapping between intensity and color similar to how the old Color and Contrast controls did in SonarWiz 5, albeit with a more sophisticated approach.
The first new color mode we are adding is a histogram matching capability. This capability is designed to automatically adjust the contrast and brightness of each trackline in a mosaic to better match the global histogram. This should be particularly useful for mosaicing sonar systems that do not have a steady state base intensity level, such as many SAS systems or other sonars doing real-time gain processing.
Figure 2A shows an example of six tracklines processed using different gain settings. Using the default linear mapping, the northern lines are much darker than the southern ones. The new Histogram Match function (Figure 2B) will compare the histogram of each track line to the global target and adjust the color mapping of each line so that they have approximately the same contrast and brightness throughout the mosaic.
The beauty of this algorithm is that it is as fast as switching drawing modes and waiting for a screen refresh. There is no need for the user to manually adjust the gains.
The second coloring mode we are adding to SonarWiz is based on histogram equalization. The Histogram equalization algorithm looks at the input backscatter values and spreads out the colormap so that each color in the palette covers approximately the same surface area in the map. Histogram equalization is useful for sonar data because the histograms of most sonar data are left-skewed. That is, most of the values of interest are down low with a very long tail towards high values. Histogram equalization spreads out the colors so that more details can be seen in the normally darker (lower intensity) parts of the data set. Compare Figure 3A to Figure 3B. There is more contrast throughout the image, especially in the softer (darker) sediments
Adding histogram equalization and histogram normalization opens the door to other image processing transformations that might prove useful for SonarWiz 8 and beyond. We’d love to hear your thoughts on these ideas. Look forward to seeing these new options in SonarWiz later this spring.