An Evaluation of the Ability of Commercial High-Resolution Multispectral Remote-Sensing Satellite Data to Identify Archaeological Resources

Summary

An evaluation of the ability of low-cost commercial satellite multispectral imagery to identify ephemeral archaeological surface sites was undertaken on the Nevada Test and Training Range in Nevada. Archaeological data from field inventories were used as a control sample. Using the QuickBird II sensor, significant differences were identified in mean reflectance in the blue, green, and near-infrared bands between on-site pixels and adjacent off-site pixels on a single site with the densest surface artifact coverage. A supervised classification using a neural network classifier with a training set based on many known sites and another based on only the dense site was unable to successfully identify and classify archaeological sites across the study area. An additional classification using a spectral curve for chert also failed to sufficiently identify archaeological sites. Although some tests indicate spectral differences may exist and be detectable on sites with dense surface artifacts, these differences are not sufficient in the study area or with existing sensors to remotely identify archaeological sites across large spatial areas.

This Resource is Part of the Following Collections

• Nellis Air Force Base Cultural Resources Archive   •
• Nellis AFB - Survey Reports

Keywords

Spatial Coverage

Individual & Institutional Roles

Contact(s): Nellis Air Force Base CRM Manager

Record Identifiers

Technical Report No.(s): 06-61

File Information