I had some trouble wrapping my head around 3D to 2D projection, but something that helped was thinking about how my own vision works. Despite living in a 3D world, everything we see is technically a 2D image. Our brain receives two sets of information set at slightly different angles, and then interprets those two images together as a single image with a sense of depth. This is similar to how the projection matrix converts the 3D coordinates of some point to 2D coordinates, taking into account factors like depth and distance.
aravmisra
@weinatalie I really appreciated the explanation and the analogy of vision. I was still having trouble intuitively grasping the parallel, so to supplement your comment I looked up how our vision worked, and found a reference to 'stereoscopic vision' here: http://scecinfo.usc.edu/geowall/stereohow.html
I found this intriguing and I wanted to learn more, so I did more research and the HIT lab at UW had some more explanation, which I found useful (perhaps the mentions on the benefits to humans/evolution in the two articles aren't as relevant, but the other stuff was definitely helpful in building intuition): https://www.hitl.washington.edu/projects/knowledge_base/virtual-worlds/EVE/III.A.1.b.StereoscopicViewing.html
colinsteidtmann
I traditionally think of the z-axis as the one that represents height. Based on the this perspective I would have thought the camera view would always be an arial shot. Are the axes pretty arbitrary? Like if we had an arial photo would we overlay the y and x axes on it the same way we're doing it to the llama?
I had some trouble wrapping my head around 3D to 2D projection, but something that helped was thinking about how my own vision works. Despite living in a 3D world, everything we see is technically a 2D image. Our brain receives two sets of information set at slightly different angles, and then interprets those two images together as a single image with a sense of depth. This is similar to how the projection matrix converts the 3D coordinates of some point to 2D coordinates, taking into account factors like depth and distance.
@weinatalie I really appreciated the explanation and the analogy of vision. I was still having trouble intuitively grasping the parallel, so to supplement your comment I looked up how our vision worked, and found a reference to 'stereoscopic vision' here: http://scecinfo.usc.edu/geowall/stereohow.html
I found this intriguing and I wanted to learn more, so I did more research and the HIT lab at UW had some more explanation, which I found useful (perhaps the mentions on the benefits to humans/evolution in the two articles aren't as relevant, but the other stuff was definitely helpful in building intuition): https://www.hitl.washington.edu/projects/knowledge_base/virtual-worlds/EVE/III.A.1.b.StereoscopicViewing.html
I traditionally think of the z-axis as the one that represents height. Based on the this perspective I would have thought the camera view would always be an arial shot. Are the axes pretty arbitrary? Like if we had an arial photo would we overlay the y and x axes on it the same way we're doing it to the llama?