My previous effort at this track I entitled the Snaefell Mountain course, which is as far as I'm aware the official title of the circuit used for the famous motorcycle races on the Isle of Man. I intended the name to distinguish it from other efforts, but probably the result was a bit of confusion as to what it was. It was built from Google Maps data and then adjusted here and there by hand, mostly using video of Mark Higgins' runs as a reference. In it's current state it looks fairly rough and ready in most places and needs a lot of detailing to make it up to something that can live with the standard of good mod tracks that are still being released for AC. But I decided to leave it as it was until there was a good reason for restarting the project. Once I got into using lidar data, thanks to Wurstkoffer's method, that's all I was interested in working on. Now, thanks to a new Manx friend of mine, I've got hold of 50cm data for the entire island. With this I plan to build a new version of the complete circuit, this time with accurate elevations, cambers, and road widths. I've finished merging the lidar tiles into a single file that covers the whole circuit. First step is to get Wurstkoffer's hack working again, and try a smallish sample to do a bit of testing...
Right now I'm looking into writing a simple plugin for RTB to read lidar data without the hack. If I convert the data to text files containing an array of height values, in theory it should be reasonably straightforward. Only two problems; although I know a bit of C, I've never used C#, so starting from scratch there. Second, the lidar source data I'm using is based on the OSGB grid reference system, and converting the lat/longs that RTB asks for is not a simple task... Any suggestions or offers of help would be greatly appreciated
Good news, converting lats/longs to OSGB turns out to be literally two lines of code, once the GeoUK package is installed. File handling will be more of a problem, since (for this project anyway) I'll be dealing with 188 data blocks, each of which is an array of 4000 by 4000 height values, anything up to 400MB in size. So far my test project takes a sample lat/long, converts it, selects the data file, reads it into an internal array and picks out the required value. But I'll need a more efficient method than doing that for every request...
