CGAL C++ Programming

I need a command line application made from the two following CGAL packages: 1. Classification: [login to view URL] 2. Point Shape Detection: [login to view URL] The application will need to build for Linux (debian 9) with CGAL version CGAL-4.13. The various options of the classification functions should be passed as command line variables in standard linux command line options. The same goes for the point set shape detection portion. You will need to read through the manual links given above to determine the options available. Here are some examples: EXAMPLE 1: pointmodeler -i STDIN -o STDOUT --classify='{ json data here }' --shapify='{ json data here }' EXAMPLE 2: pointmodeler -i [login to view URL] -o [login to view URL] --classify='{ json data here }' The options for each "sub command" (classify or shapify) should take JSON format. If both commands are specified as in "--shapify" and "--classify" a header will be used so the change in output from classified points to shapified points will be dellineated by the header. EXAMPLE LIDAR POINT HEADER: "X","Y","Z","Intensity","ReturnNumber","NumberOfReturns","ScanDirectionFlag","EdgeOfFlightLine","Classification","ScanAngleRank","UserData", "ID" As far as the output header format for the 3D format, I am leaving that to you but clarify further below. INPUT: The format the app will accept is included as an attachment. [login to view URL] There is just one additional column, that is "id" which is where this application differs from the las2txt format. When running the classify portion of the application, it will simply update the classify column of data. Much of this process is outlined in this doc: [login to view URL] In additional to the manual links given at the beginning of this document. Once points are classified, shape detection should be ran. The output of "shapify" will be the simplified geometry types. For example, if a house is made of up of 8 planes once it has been converted from points to a simplified 3D object, that object should be marked as 1. building 2. only return the simplified 3d objects and not the individual point clouds The output should match what is shown in that picture of page 3. 3. One object set should be distinguishable from another For example, all 3D objects that make a single home should be tied together together in a way that when they are output, this is distinguishable in the data. The original input may have consisted of more than 80Megabytes of points and output a file as small as 2Megabytes due to the geometrical simplifications. The output should not be overly simplified but also be controllable through the options with simplification for each different classification uniquely controllable. For example, these would be options passed as JSON to the shapify command: { buildings: { simplicity: 4, // from 1-10 would determine how aggressive the simplification would be more_options: 2.5, ... }, vegetation: { simplicity: 7, // from 1-10 would determine how aggressive the simplification would be more_options: 5.5, ... }, ... } The output should be a format of the 3d objects should be compatible with postgis type geometries into either JSON, text file or STDOUT. [login to view URL] One of the most crucial portions of the app is that it must have accurate roof edges, angles and building edges as this will be mapped to aerial imagery. I have attached a sample input csv file.