Hello.
The original code performs two registrations at different scales (5 and 1) but uses the same initial pose estimation for both:
transformation, _ = self.registration_at_scale(scan_tobe_mapped, global_map_in_FOV, initial=pose_estimation, scale=5)
transformation, fitness = self.registration_at_scale(scan_tobe_mapped, global_map_in_FOV, initial=pose_estimation, scale=1)
I think you want to put the rough result to the initial of the precise icp, so maybe:
transformation, _ = self.registration_at_scale(scan_tobe_mapped, global_map_in_FOV, initial=pose_estimation, scale=5)
transformation, fitness = self.registration_at_scale(scan_tobe_mapped, global_map_in_FOV, initial=transformation, scale=1)
Hello.
The original code performs two registrations at different scales (5 and 1) but uses the same initial pose estimation for both:
transformation, _ = self.registration_at_scale(scan_tobe_mapped, global_map_in_FOV, initial=pose_estimation, scale=5)
transformation, fitness = self.registration_at_scale(scan_tobe_mapped, global_map_in_FOV, initial=pose_estimation, scale=1)
I think you want to put the rough result to the initial of the precise icp, so maybe:
transformation, _ = self.registration_at_scale(scan_tobe_mapped, global_map_in_FOV, initial=pose_estimation, scale=5)
transformation, fitness = self.registration_at_scale(scan_tobe_mapped, global_map_in_FOV, initial=transformation, scale=1)