Event

Developer calibrates wrist-mounted iPhone and iToF sensor using ECO65 robotic arm

1 reports · 1 sources

A developer (Reddit user No_Quail6685) used the motion of a Realman ECO65 robotic arm as a calibration target to align a wrist-mounted iPhone with an external iToF sensor, and imported RGB color-depth point clouds and joint motion data into an office digital twin. The iToF calibration approach: average 10 frames per pose, isolate pixels that change when the robotic arm moves, then fit these points to the robot model using ICP. The developer reports a median fitting error of 7 mm, but emphasizes that this is a fitting residual, not an independent accuracy measurement; visible motion of the J3 and J4 joints was insufficient and provided limited help. iPhone depth estimation results remain unreliable: only 6 out of 20 frames passed inspection.