An in-depth exploration of how machine studying stacks up in opposition to conventional coastal erosion monitoring strategies
Deep studying (DL) is the one option to clear up the issue. That’s the implicit assumption in a lot of the analysis I learn. I’m all the time tempted to agree. However, that’s in all probability as a result of my PhD can be ineffective in any other case.
Fortunately, the extra I learn, the extra I realise that distant sensing is stuffed with issues the place machine studying may help. These embrace monitoring air high quality, estimating soil moisture, assessing crop well being and monitoring pure disasters. It’s also true for my space of analysis — monitoring coastal erosion.
The shoreline is lengthy! The size means we have to automate some duties to successfully monitor all of it. On the similar time, noise attributable to components like land improvement, cloud cowl and offshore winds and swells means conventional, deterministic strategies can fail. It’s dealing with variations like this the place machine studying thrives.
Deep studying, a subfield of machine studying, has emerged as a priceless device in distant sensing, providing options to unprecedented challenges and creating new alternatives in distant sensing functions,