I'm new to computer vision. I'm trying to solve a problem that I find interesting. My goal is take a 2 pictures of an object, that is able to move freely (like a ball). then compare the two pictures, to see if the balls elevation has risen. I've looked into examples of finding distance/hight but, it was always with an object with fixed distance. I've looked into projective geometry/other programing libraries but, i've just ended up more confused than finding some direction. I'm not sure what type of resources I need to be looking at. Any advice would be great!
Asked
Active
Viewed 106 times
0
-
1What do you mean by distance/hight ? – Sep 28 '18 at 07:58
-
If the object is able to move how can I find the distance, and if it moves up how can I find how high it moved – Joshua Jones Sep 28 '18 at 11:07
-
Distance in the image (in pixels) or real distance (in meters or similar)? – Nico Schertler Sep 28 '18 at 14:40
-
Sorry, real distance! @NicoSchertler – Joshua Jones Sep 28 '18 at 14:52
-
That's not as easy as it might sound. You would need to estimate the depth of the ball (e.g., from its size) and then use the intrinsic camera matrix (that you need to calibrate) to transform the pixel position to a real-world position. – Nico Schertler Sep 28 '18 at 15:02
-
@NicoSchertler If we have known starting point where we know the distance. Then the ball moves. Would that be more easier? – Joshua Jones Sep 28 '18 at 15:20
-
I guess you could estimate the depth from the ratio of the sizes of the current frame and the reference frame. Not sure how accurate this will be, though. But you still need a calibrated camera. – Nico Schertler Sep 28 '18 at 18:49