  The timestacks can reveal interesting features, such as decreasing wave speed in shallow water. Lines in the timestack correspond to velocities. An oceanographer seeing the timestack to the right would immediately recognize the lines with increasing slope. These correspond to waves slowing down as the water shallows; a well-known phenomenon. Here is an early timestack from Mokuleia showing the swash. It also illustrates the necessity for better data acquisition (gaps and repeats).  The next timestack is from Waimea. It shows a large wave, and the effect of changing cloud cover. The auto-exposure was able to cope. If you zoom in closely, the pink dots shows the result of the swash detection algorithm. The instantaneous high-water line is tracked reasonably well.  The human eye and cognitive system is excellent at tracing the outline of the swash, which would then yield a data product related to wave frequency and amplitude. For a computer to automate the same task, a detection algorithm is needed. I tried the majority of common edge-detection algorithms with disappointing results for this application. Most of them rely on spatial differences; which tend to amplify noise. The next image shows a typical result from a common edge detection algorithm. Notice that the image has been reduced to binary logic (0 or 1), from which a position is extracted. False positives and negatives in the binary image cause error and biases.  |