# Computer Vision Distance Measurement node?

#1

Hi guys.

is there any Computer Vision Distance Measurement nodes or tips?
I need to distance camera to small object.
so kinect and other depth camera is bit rough for me.

thx

#2

i fear there isnt a single node solution. you need to know the intrinsic parameters of your camera. like focal length, size of sensor so on… you may need to calc these params by chessboard calibration if not provided in specs. then you need to know the real height of your object and the height of the object in the captured image (this could be achieved by some kind of segmentaton) followed by some math like:
$\text{Distance&space;to&space;object}(mm)&space;=&space;\frac{f(mm)\times\text{real&space;height}(mm)\times{\text{image&space;height}(pixels)}}{\text{object&space;height}(pixels)\times\text{sensor&space;height}(mm)}$ -
this approach will only work for straight perspective. if you thinking about moving your camera around an arbitrary 3d object this will not work. you should go for depth sensor then (where you also wll need some kind of segmentaion to identify your object). or use an AR Library which allows to detect 3d features.
best, m.

#3

ah and as opencv3 introduced object tracker you could either use its c# wrapper or check the progress of the awesome https://github.com/vvvv/VL.OpenCV contribution.

#4

Thank you very much Milo.

I fear there isn’t a single node solution.

OK i understand. I’m also can’t find a single node on this website.

opencv3 and VL.OpenCV contribution.

Thank you also here. Let me check

Actually My project is not art. Making vegetable harvest machine.
I want to measure distance of the robot arm finger and vegetable. for compensating for position of the arm.
Measure distance and feedback it to the robot-arm control system is my goal.

thx

#5

you may could use something like this which also would be even more precisely.

closed #6

This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.