Preliminary Image-Observability Screening of Human-Interpreted Parcel Boundaries Using Radiometric Edge Proximity: A Case Study in Vientiane, Lao PDR

Reliable cadastral modernization requires distinguishing legally authoritative boundaries, human-interpreted parcel geometry, and image-visible evidence. This study examines the spatial proximity between a human-interpreted parcel boundary layer and unfiltered radiometric edge evidence derived from the same high-resolution orthophoto in Vientiane, Lao PDR. The analysis is not a cadastral accuracy validation and does not treat Canny-derived edges as an independent or higher-accuracy reference. Instead, it quantifies parcel-level boundary-to-radiometric-edge offset as a preliminary image-observability screening layer. For 89,763 parcels, nearest-edge offsets were summarized using mean, median, and upper-tail metrics and examined through spatial clustering, Canny threshold sensitivity testing, and a sample-based visual audit. Across parcels, mean offset values averaged 0.72 m, while median-offset values had a median of 0.21 m. Sensitivity testing showed that absolute offset magnitudes vary with edge-detection thresholds, indicating that the metric should not be interpreted as ISO-style positional accuracy. The mapped clusters, therefore, indicate an elevated boundary-to-radiometric-edge offset, as opposed to a confirmed cadastral error. The workflow is intended to support preliminary prioritization for expert visual review, semantic filtering, cadastral record checking, or field verification.