*8.2. Outlook*

The determination of the parameter *MTD* using 3D point clouds could be performed in a similar way like the parameter *Ra*, with the difference that the reference plane is placed on the uppermost points of the point cloud. In addition, in this case, the distances between the points and the plane have to be summed and divided by the total number of points. This calculation procedure principally corresponds to the determination of *MTD* as derived by the sand patch method. In practice, though, the implementation of a calculation procedure for *MTD* holds some difficulties. The point clouds reconstructed by our measurement system usually consist of several million points, some of which, inevitably, are outliers. Therefore, choosing which points to use when defining the *MTD* reference plane would be difficult.

In the future, however, our measurement system enables novel opportunities for the investigation of technical surfaces. Many roughness parameters are designed for lines and are not suitable for the description of entire surfaces. Hence, besides the adaptation of further roughness parameters (e.g., *MTD*, *Rv*, *Rp*, *Rt*, Abbott-Firestone curve) to 3D point clouds, we think about designing new roughness parameters, which represent the surface properties in a better way. In particular, for different use-cases, different parameters should be considered. For example, a distinction has to be made between roughness parameters used to estimate the amount of coating material required to cover the surface and roughness in terms of adhesive bond of a surface. In this context, for example, the investigation of the gradients in the point clouds would be feasible.

The roughness of a 3D reconstructed surface essentially depends on the measurement resolution. For a resolution-independent estimation, the fractal dimension should be considered, as, for example, presented in [19].

As a further outlook, additional empirical studies, including comparisons with different methods (e.g., laser triangulation, sand patch method) have to be done in order to validate the camera-based system and present the method applicability. Therefore, as a first step, the measurement system could prove to be a supplementary measurement system to the sand patch method and after getting approval, it could perhaps become an alternative method.

Current high-end smartphones are equipped with an integrated graphics processor (IGP), which also enables parallel computing. In particular, smartphones with the Android operating-system provide through RenderScript [52] a powerful API for implementing parallel algorithms which can further be executed by the IGP. Therefore, it is conceivable that our image-based measurement method, as presented in this article, can be adapted onto mobile systems as smartphones since they provide all necessary components, such as, for example, a high-resolution camera and powerful processing unit. In addition, since nowadays almost everybody owns a—more or less powerful—smartphone no additional hardware would be necessary.

**Author Contributions:** B.Ö., R.S., and J.B. worked out the idea and designed the concept; B.Ö. designed and implemented the software and performed the experiments; R.S. and J.B. administrated and supervised the research project; B.Ö. wrote the paper, R.S. and J.B. reviewed the text and offered valuable suggestions for improving the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Bundesministerium für Wirtschaft und Energie, gran<sup>t</sup> number 16KN062126.

**Informed Consent Statement:** "Not applicable" for studies not involving humans or animals.

**Data Availability Statement:** "Not applicable" for studies not involving humans.

**Acknowledgments:** The authors acknowledge the company of Nedo GmbH & Co. KG for the excellent collaboration in the research project. The authors would further like to thank the Institute of Building Materials Research (ibac) of the RWTH Aachen University for providing the concrete specimens and corresponding reference values. Additionally, the authors would like to thank Louis Makiello for proofreading.

**Conflicts of Interest:** The authors declare no conflict of interest.
