@nadar wonders if there is a way to improve the speed of the surface detection process. @moon will have a look and implement multi-threaded version of the process.