HALCON’s surface-based matching is a powerful, field-proven technique to locate individual objects within point-clouds. Translating the found poses into robot coordinates requires hand-eye calibration. If your 3D sensor also produces a 2D image (like a stereo camera system, for example), you can apply a standard calibration procedure using a dot-pattern. Instead of a 2D pattern, however, you can also use any 3D object as a calibration target, and determine its pose with find_surface_model.


You can feed the found pose into the calibration model with set_calib_data_observ_pose which effectively replaces find_calib_object. The two figures below show the workflow.