now studying machine learning.
ITmedia�̓A�C�e�B���f�B�A�������Ђ̓o�^���W�ł��B
。关于这个话题，新收录的资料提供了深入分析
Что думаешь? Оцени!
The solution: good old Perlin noise. A global noise field determines tree density and building placement, completely separate from WFC. Areas where the noise is above a threshold get trees; slightly different noise drives buildings. This gives you organic clustering — forests, clearings, villages — that WFC could never produce. I also used some additional logic to place buildings at the end of roads, ports and windmills on coasts, henges on hilltops etc.