A statistics based rigid body transformation approach for precise georeferencing of cadastral parcels preserving shape without stretch

Positional accuracy is critical in cadastral mapping, but achieving it through georeferencing without distorting a parcel’s shape and dimensions is equally crucial due to legal and financial implications. Conventional polynomial-based georeferencing and rubber sheeting reliably improve accuracy by scaling and stretching parcel geometry. Built-in similarity transformation tools solve simultaneously for translation, rotation, and scale; unless scale is constrained to unity, the fit drifts, subtly rescaling every parcel. Both require manually digitizing parcel-to-survey-point links across hundreds of point pairs. CAD-based alternatives bring data conversion challenges. This study presents a rigid-body approach that preserves parcel geometry while achieving the required accuracy. Histograms of outlier-filtered spatial discrepancies in offset and direction revealed a dominant mean, applied iteratively to the whole layer as a single rigid body until the required accuracy was reached. The approach, applied to 180,000 parcels around Riyadh, is evaluated here on a representative sample of 15 neighborhoods (19,012 parcels) against 725 survey points, achieving an NSSDA positional accuracy of 0.244 m at the 95% confidence level, within the 0.25 m requirement, while preserving every parcel’s shape and area. Affine and Similarity transformations achieved 0.211 m and 0.218 m, also within the requirement, but distorted every parcel’s area by up to 49.3 m 2 and 60.3 m 2 ; rubber sheeting, though fitting control points exactly by construction, distorted areas by up to 3521 m 2 . The proposed approach trades about 0.03 m of accuracy for exact preservation of parcel geometry, and suits cadastral programs where parcel dimensions must not be altered.