This paper provides a preliminary analysis of reconstructions of seven disaster props from datasets collected over five flights with a small unmanned aerial vehicle and finds 181 anomalies present in four categories. Commercial and free software are available to create reconstructions from digital images, but their accuracy and limitations have not been formally explored for disaster assessment. This work analyzes the output of two packages, Agisoft Photoscan, and Microsoft ICE for five image datasets of seven Disaster City®props. The 181 anomalous results are divided into four categories: ghosting, misalignment, misproportions, and the Dali effect. These metrics are valuable to Urban Search and Rescue teams, structural specialists, and insurance claims adjusters who need to reliably assess whether it is safe to enter a building, or accurately estimate the severity of damage. The analysis suggests the reconstructions might be misleading and further work must be done to determine why and suggest methods of improvement.