3.2.3. Automatic Boundary Extraction Methods

The method used for the current study is based on the work in [48] and shown in Figure 9. It supports the delineation of boundaries by automatically retrieving information from RGB data that is then used to guide an interactive delineation. It consists of three parts: (a) image segmentation, (b) boundary classification, and (c) interactive delineation. The source code is publically available [49].


**Figure 9.** Boundary mapping method: (**a**) Multiresolution combinatorial grouping (MCG) image segmentation. (**b**) boundary classification that requires line labeling into "boundary" and "not boundary" for training. The labeled lines are used together with line-based features to train a Random Forest classifier that generates boundary likelihoods for testing. (**c**) interactive delineation guided by a QGIS plugin.

An analysis on our study area in Kajiado assessed the ABE application for extracting visible cadastral boundaries using the acquired UAV images [24]. During the workshops, we used the three tiles of 300 × 300 m shown in Figure 7 to demonstrate the boundary mapping method. In this way, we related the method to our conceptual framework (Figure 2). This allowed identifying and understanding bottlenecks: (i) the operational analysis questions when and why the method works better or worse compared to manual delineation and (ii) the feedback analysis investigates the method based on surveying stakeholders responses. The feedback analysis was based on discussing the strengths, weaknesses, opportunities, and threats (SWOT) of our proposed method compared to manual delineation as identified by the workshop participants.
