- Home /
Question by
Trigonaut · Feb 17, 2017 at 08:20 AM ·
2draycastingnullimages
Filling blind spots in raycast cone?
I'm having trouble making a sort of AI camera. I want to convert the view of a character into a set of data points, to be sent to a neural network with a limited number of discrete inputs (about 6). I've tried Raycast2D, and it works very well, except for the fact that small objects may hide in blind spots between rays.
Is there another way to do this, like maybe using a RenderTexture?
In the end, I'd like to basically create a list representing a 1-dimensional image for the character's view, which defines the contents of each pixel from some data retrieved from the object, like distance or color.
halp.png
(37.5 kB)
Comment