**5. Conclusions**

The overall aim of this study was to provide an UAV-based cadastral mapping workflow based on the ENVI FX module for automatic detection of visible boundaries. The study first investigated, which processing steps are required for a cadastral mapping workflow following the potential and limitations of the ENVI FX for automatic visible boundary detection and extraction.

The results showed that more correct visible object boundaries, suitable for the interpretation of land cover (cadastral) boundaries, were extracted at larger values of GSD. In addition, the identified optimal scale and merge levels for detection and extraction of visible cadastral boundaries were between 70 and 80 and 95 and 99, respectively. The identification of the optimal parameters for cadastral mapping was based on the defined minimum object area and the total number of objects from the reference data using the so-called filtering approach. The filtering approach contributed to the increased correctness of automatically extracted boundaries. The best results were recorded at the resampled UAV orthoimage with a GSD of 50 cm, and the percentage of correctness indicated was 77%, while for the completeness it was 67%. It must be emphasized that the workflow developed is applicable mostly for rural areas where the number of visible boundaries is higher compared to complex urban areas.

The workflow can be used in developing countries to accelerate and facilitate the creation of cadastral maps aiming to formalize a land tenure system and guarantee legal security to land rights holders. In developed countries, the extracted visible boundaries based on this workflow might be used for efficient revision of existing cadastral maps. However, in both cases, the extracted visible boundaries have to be validated by landowners and other beneficiaries. The extraction of visible objects can be considered as only one step in the facilitation of cadastral mapping, as extracting these is not enough for complete and correct cadastral mapping. It is worth highlighting that cadastral boundaries may, in fact, be completely inside the property and that some boundaries between properties may not be clearly visible. In order to use the proposed workflow in the cadastral domain, the approach can be expanded. Additional steps should focus on methods for the possible involvement of current

landowners in the process of cadastral mapping. The extension of the current workflow is one of the aims of the authors' further research.

**Author Contributions:** Conceptualization, B.F. and A.L.; methodology, B.F., K.O. and A.L.; software, B.F.; validation, B.F., K.O. and A.L.; formal analysis and data processing, B.F.; investigation, B.F.; resources, B.F., A.L., K.O. and M.K.F.; writing—original draft preparation, B.F.; writing—review and editing, B.F., A.L., K.O. and M.K.F.; visualization, B.F.; supervision, A.L.

**Funding:** The authors acknowledge the financial support of the Slovenian Research Agency (research core funding No. P2-0406 Earth observation and geoinformatics).

**Acknowledgments:** We acknowledge Klemen Kozmus Trajkovski and Albin Mencin for capturing the UAV data and for the technical support during the fieldwork.

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