Moisture Content Estimation of Porous Building Stones using Hyperspectral Imaging.

Area of Science:

• Building materials science
• Conservation science
• Geophysics

Background:

• Moisture is a primary driver of built heritage deterioration, causing damage through frost, salt crystallization, and biological growth.
• Accurate assessment of stone moisture content is crucial for effective heritage conservation strategies.

Purpose of the Study:

• To validate the use of shortwave infrared (SWIR) hyperspectral imaging for non-destructively determining stone moisture content.
• To create and validate moisture maps illustrating water distribution in various stone types.

Main Methods:

• Preparation of stone samples (cubic and cylindrical) from six types (Brick, Euville, Massangis, Neubrunn, Obernkirchen, Savonnières) with controlled moisture levels.
• Acquisition of SWIR hyperspectral images and processing using the Normalized Relative Arc Lengths (NRAL) method to generate moisture maps.
• Validation of moisture maps through petrographic examination and quantitative comparison with gravimetric moisture content.

Main Results:

• Strong agreement was found between moisture content estimated via SWIR imaging and gravimetric measurements.
• Root mean square errors for moisture content estimation ranged from 1 to 2 g/g × 100.
• Moisture maps accurately depicted water distribution, correlating with stone pore characteristics and grain arrangement.

Conclusions:

• SWIR hyperspectral imaging is a reliable non-destructive technique for assessing stone moisture content in built heritage.
• This method provides valuable data for understanding moisture-related deterioration and informing conservation efforts.
• The NRAL method effectively generates moisture distribution maps for diverse stone materials.