2

I was trying to practice finding vanishing points in different images, and most tuts I followed showed images with at least some rigid surfaces like a table or something, but, I found a particular image, that I think has only one vanishing point, but am not quite sure how to plot the vanishing point of. How do you suggest I plot vanishing points for such images?

EDIT: I need to find the vanishing point and the horizon line because, in blender there is a tool that allows you to set the focal length of the camera feeding in these two data, and I'm unable to do so in the specific cases of such images as below:

image

4

Nope. There is not only one vanishing point. There is an infinite number of them.

You are thinking in terms of a cityscape where we have orthogonal walls. Each wall has its own set of vanishing points.

But when you have a natural organic landscape you do not have clear angles, so forget about them. That is why you do not see clear vanishing points.

A segment of the ice has one until it changes direction, until it changes size, until... it is better not to think on segmented vanishing points at all...


Edited:

Let's imagine there are some towns on the horizon. If I choose to build a road for them, I would build a road with parallel edges. But as you can see I can build different roads, each with its own perspective.

The yellow and the red lead to the same town, but they have a different perspective. The red one looks as if the camera is lower on the ground.

The orange and the blue leads to different towns on the horizon. And all of them could be feasable.

enter image description here

We are used for a two-point perspective because we live in a constructed world, A room, a table, a city block.

So let's try to put a rectangular base, probably for a small building.

The (A) image does look odd. Let's find out why.

We are assuming that the horizon is flat. If I draw some perspective lines to the horizon and one of the 90° border (closer to the camera) we can see that the other two are off (B). We simply need to correct them (C)

enter image description here

But this does not mean that they are the only vanishing points.

Here is a different perspective with different vanishing points with different "focal length" and the plane is really on the ground based on its own arbitrary vanishing points.

enter image description here

So. There is not only one vanishing point. There are endless ones. Choose two and enjoy!


Complementary post: At what point 1 point perspective become 2 point perspective?

  • you see I need the vanishing points and a horizon the adjust the focal length of the camera in blender to perform camera mapping, so I was trying to find it..... is it really impossible? – mathmaniage Oct 13 '18 at 16:51
  • I don't know that "organic" has anything to do with it - rocks are inorganic but similarly don't have a single vanishing point – Zach Saucier Oct 13 '18 at 16:57
  • @mathmaniage please add this info to the question. I will expand the specific case on my answer later (When I arrive home) – Rafael Oct 13 '18 at 17:04
  • @Rafael , I've added the information. – mathmaniage Oct 14 '18 at 2:53
  • @Rafael , here's how it's usually done: youtube.com/watch?v=yeWzQYh7iH0, but, landscapes no clue – mathmaniage Oct 14 '18 at 2:57
2

Most of us can see it and mathematicians have also proven that in a perfect 2D perspective image of a 3D scene any set of parallel lines seem to point towards a single point which is called the vanishing point.

Every set of parallel lines have their own vanishing point. That point depends also on the used imaging method and its parameters.

Vanishing points are useful tools for artists when they construct a drawing which has ordinary buildings and straight roads. That environment in 3D has plenty of parallel line sets. Vanishing points help to draw those parallel line sets, as you obviously have learned.

Trying to figure out vanishing points for something else than sets of parallel lines, is mathematically nonsense. If an artistically oriented person some day tells that he can see the vanishing points also for curved forms, I do not claim he is lying or he has taken something. His own experience can bypass the limits what's defined exactly in math. Unfortunately we have no language to describe experiences exactly.

  • I agree with @user287001, but would add: for some curve sequences which are inter-related in natural landscape images (contours, river banks, beaches etc) you can derive implied parallel lines which run tangent to those curves, and in averaging those, you can end up with some generalised vanishing points - but it's messy and imprecise perforce, and most frequently is of limited utility. For OP's specific use, I would draw onto a copy of the image some linear elements by hand, using artistic judgement, use that in Blender's photomatch, and replace image afterwards. – GerardFalla Oct 15 '18 at 16:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.