M3dmetFit

This function fits a 3D geometry to a point cloud or depth map using an iterative fit operation. The function starts from an initial fit estimate of the size and position of the 3D geometry. The function then internally iterates through better estimates of the correct size and position of the 3D geometry until a specified stop condition is reached. Upon each iteration, the function calculates the average root-mean-square (RMS) error per inlier, until a stop condition is met.

The initial fit estimate is calculated according to the mode specified using M3dmetControl() with M_ESTIMATION_MODE. A set of inlier points is then defined according to OutlierDistance, and a new size and placement for the 3D geometry, which reduces the average RMS error per inlier, is calculated. This process then repeats until a stop condition is met.

If the initial fit estimate mode is set to M_FROM_GEOMETRY, the fit operation will use the specified geometry in the fit 3D metrology context. By default, the context contains an undefined geometry. If you try to initiate the fit operation and use the undefined geometry to determine an initial fit estimate, an error will occur. You can use M3dmetCopy() to copy a defined 3D geometry object into the fit 3D metrology context.

If you perform a fit operation using a depth map, it is possible to specify a ROI in raster format, where only the values inside the ROI will be used in the fit. Note that missing data (or invalid points) in the depth map (or point cloud) are ignored when fitting.

When fitting, it is also possible to specify an outlier distance, beyond which no input data is considered for the fit operation.

After calling M3dmetFit(), you can inquire whether the operation was successful using M3dmetGetResult() with M_STATUS, along with several other metrics. To retrieve the calculated best fit 3D geometry or the inliers, use M3dmetCopyResult().

Note that the fit operation does not support 3D box geometry objects. All other 3D geometry objects are supported.