Hmmm.... not sure about this. Mapping seagrass is not really up my alley, but I am skeptical about using such images since they are just RGB images. I would imagine that you could make some sort of crude mask to isolate the seagrass by defining some sort of threshold for RGB values.
For example, every pixel has 3 values associated with it: red, green, and blue. Suppose you decide that the seagrass in your image seems to range between 100-255 for green, and red and blue values are each 0 (just an example). You could then make a simple program (or use software) to display the seagrass by only including the pixels that meet your criteria (i.e. green values between 100-255) and mask out the rest that isn't seagrass. The trick is coming up with a way to distinguish the seagrass from any other objects in the image. Let me know if that makes sense.
Professor Gilerson at the City College of New York would know a lot more about this than me. You could try searching Google Scholar for papers that he has authored--- perhaps that would give you some insight. Cool images by the way!