Using UAV-borne remote sensing to develop a classification method for footpath degradation from path mapping and degradation characterisation

Abstract

Increasing recreational touring in natural areas is putting increasing strain on footpath infrastructure. The small-scale and dispersed nature of footpath networks makes assessing and monitoring degradation a significant challenge for land managers. This study utilised high-resolution UAV imagery to identify footpath locations from object-based image analysis and supervised classification, assessing the strength of varying segmentation and classification approaches.

Path degradation was then mapped to 0.5m sections of path, using degradation indicators of width, incision, braiding, and rugosity. A two-layer segmentation, with a Bayesian classifier, yielded the most accurate path footprints, and degradation was accurately determined, using a degradation indicator ruleset. This methodology presents a feasible option for land managers to develop a path network inventory and improve the efficient allocation of resources for path maintenance by identifying which areas require the most attention.