New approach to accuracy verification of 3D surface models: An analysis of point cloud coordinates

The precision of two types of surface digitization devices, i.e., a contact probe scanner and an optical scanner, and the trueness of two types of stone replicas, i.e., one without an imaging powder (SR/NP) and one with an imaging powder (SR/P), were evaluated using a computer-aided analysis.

A master die was fabricated from stainless steel. Ten impressions were taken, and ten stone replicas were prepared from Type IV stone (Fujirock EP, GC, Leuven, Belgium). The precision of two types of scanners was analyzed using the root mean square (RMS), measurement error (ME), and limits of agreement (LoA) at each coordinate. The trueness of the stone replicas was evaluated using the total deviation. A Student's t-test was applied to compare the discrepancies between the CAD-reference-models of the master die (m-CRM) and point clouds for the two types of stone replicas (α=.05).

The RMS values for the precision were 1.58, 1.28, and 0.98μm along the x-, y-, and z-axes in the contact probe scanner and 1.97, 1.32, and 1.33μm along the x-, y-, and z-axes in the optical scanner, respectively. A comparison with m-CRM revealed a trueness of 7.10μm for SR/NP and 8.65μm for SR/P.

The precision at each coordinate (x-, y-, and z-axes) was revealed to be higher than the one assessed in the previous method (overall offset differences). A comparison between the m-CRM and 3D surface models of the stone replicas revealed a greater dimensional change in SR/P than in SR/NP.