Binocular distance measurement, the requirement is specified in the space of a point distance is ok (I was in the white box drew a black dot on a Angle), by matching the feature points now, around the image pixel coordinates of the point at which have been identified, triangulation distance Z=f B/d, the f is the focal length, B is the baseline distance, d is the parallax,
The question is: parallax for pixel difference, the unit is the pixel, pixels and f and B is the length of the unit, so that the distance Z is not a distance between the calculated ah... Focal length is formula of focal length f with pixel fx, fy?
O god,
Binocular calibration (using MATLAB to get the space diagram shows two nearly parallel to the camera, stereo calibration, so I didn't want to directly to the measurement of matching feature points, then using the triangulation method to calculate distance)

CodePudding user response:

I've never used this method, you alone formula, you this is
Is distance Z (physical distance)/focal length (pixel)=baseline distance B (physical distance)/inspection (pixel)
If formula was established,
Instead you get Z unit of distance, physical distance,

CodePudding user response:

Z=f/B x1 - x2 I understand it, the formula of x1, x2 should be image coordinate system, rather than pixel coordinate system, and I want to ask the landlord to make come not, generally we all heavy projection matrix calculation using the funnel type model Z, the parallel model precision can make?

CodePudding user response:

Z=f B/d
Z coordinate=focal length * baseline/horizontal parallax
Z, B, m for the distance unit, mm, etc.;
F, d unit for m, mm, etc., or pixel, but must be consistent,
Transformation of two classes of units by size, like yuan the pixel size, generally 1 pixel for camera chip to the physical size (microns),
If two cameras are not parallel, namely the relationship about translation, the parameter is not rule is, not directly with the formula, need transformation;
Like after rotating translational elliptical can no longer use standard elliptic equations, said

CodePudding user response:

say first conclusion: focal length f is pixel fx, fy
Fx and fy is usually of the same or similar, which is a specific, depends on your algorithm in pixel coordinates and camera coordinates the relationship between the

CodePudding user response:

silly don't lack reference 4 floor response:

say first conclusion: focal length f is pixel fx, fy
Fx and fy is usually of the same or similar, which is a specific, depends on your algorithm in pixel coordinates and camera coordinates the relationship between the

Ok, thank you ~

CodePudding user response:

Ask a question binocular distance measurement stereo calibration images need to be cut

CodePudding user response:

Triangulation method used is the premise of the two cameras must do rectification, the two camera rotation to a horizontal plane and the distortion

CodePudding user response:

Could you tell me the distance also need two camera images of the stereo matching? Use SGBM or bm, will appear in the camera around the problem of matching two points less than? Because I see other bloggers have parallax figure, is not very complete,