I have a problem with understanding how to achieve good results (for the eye) when applying texture on a terrain mesh. I know what U-V coordinates are and I know how to apply texture to mesh.
It is hard for me to explain what I do not understand

which explains how a terrain and textures has been created.
Firstly we have a terrain:

And then we have applied texture:

On the second image you can see the line where grass meets with sand. This line is very irregular, it looks natural. However the underlying mesh does not have its veritces at the exact place where the "grass line" is visible. Does it mean that the whole texture has been precomputed in memory (combined from few textures by texture splatting) and then, this one single texture has been applied to the mesh? Article says that the heightmap for terrain is 513x513 so I assume that the used texture is 512x512. So texture with this size has been stretched on a terrain with size 10x10 km.
Is my assumption correct and this is how its done?