Distance between the two points

[SCHAR9]

Hello,

As part of a project, I was interested in the F200 camera and its associated SDK that seems to suit my needs.

I need some light to be shed on the technical possibilities of this camera.

For example if I select two points on an object facing the camera, is the camera able to provide me the distance between the two points as well as the distance of each with the camera and with what precision?

Thank you!

[MartyG]

Here are some specifications for the F200.

In regard to your question: I believe that what you are seeking is possible but there are people on this forum who can provide the specific details of how to do so better than I can. Hopefully they'll chip in with advice later on when they see this.

[jb455]

You'll have to calculate the distance yourself (easy once you have the points, search this forum for more help if you need it). In our testing we usually get accuracy to less than a millimetre at a distance of about 30cm from the camera. The main source of inaccuracies we've had is selecting the right points to measure between (fingers aren't particularly precise!)

[SCHAR9]

Thanks for the quick answers.

I'll take an example to be more accurate on what I want.

Let's take a cube, if we put some colored spots on each corners, with this camera, can we get the distance between two spots placed on the same face regardless how the cube is turned ? Will the depth be taken in account in the calculation or can we dispose of the depth of each points to do the calculation by ourselves ?

[jb455]

It will still work if you rotate the cube, as long as both cameras can still see the two points. The accuracy may drop off though if the face is at a shallow angle.

Not entirely sure what you mean in your second question. To calculate the distance between two points, we get (x,y,z) "world coordinates" from the camera which tells us where each point is in real space, where the x & y coordinates are the distance in mm from a central point in the x-y plane, and the z coordinate is the distance in mm of that plane to the camera.

[SCHAR9]

Understood. For us, the keywords were "real space", it suits us that the world coordinate are given in mm, this is what we were searching for. Now, for the precision, what will be the accuracy (one mm, two ?) when one face of the cube is mostly facing the camera (we disregard the aberrant cases, only the optimal ones are taken into account) ?

Thanks again.

[jb455]

As I said before, we usually see accuracy of a millimetre or less when testing (measuring distances of up to 30cm at a range of 30-50cm) assuming you pick the right points to measure between. In practice, the camera performance will vary with the ambient lighting conditions and the material it's looking at - if there's too much ambient infrared around (eg, near a window on a sunny day) or the material doesn't reflect enough IR (or too much!) you'll get different levels of accuracy. Best thing to do would be to test in whatever your normal conditions will be yourselves!

[SCHAR9]

Thanks for all your advices and infos. As you said, we will now perform some tests now to see if the results are accurate enough in our normal conditions. Thanks again.