Rapid determination of pit mud moisture content using hyperspectral imaging

Area of Science:

• Agricultural Engineering
• Food Science
• Spectroscopy

Background:

• Pit mud moisture content is critical for liquor brewing, affecting aging and structural integrity.
• Traditional moisture detection methods are often slow, destructive, or impractical for in-situ monitoring.
• Developing rapid, nondestructive techniques is essential for optimizing fermentation processes.

Purpose of the Study:

• To investigate hyperspectral imaging for rapid and nondestructive moisture detection in pit mud.
• To compare modeling efficiencies in visible and near-infrared spectral regions.
• To analyze pit mud moisture distribution patterns and their implications for liquor fermentation.

Main Methods:

• Hyperspectral imaging was employed across visible (400-1,000 nm) and near-infrared (900-1,700 nm) regions.
• Various data processing techniques were evaluated for model development.
• The optimal model (SNV-SPA-SVM) was identified in the near-infrared spectroscopy region.

Main Results:

• The SNV-SPA-SVM model achieved high accuracy, with R² of .9953 and RMSEP of 0.0029.
• Moisture content was generally lower in new cellars compared to old ones.
• Moisture distribution was more even in old pit mud, and content increased with depth.

Conclusions:

• Hyperspectral imaging provides effective online monitoring of pit mud moisture for liquor brewing.
• This technology offers strong technical support for quality control in solid-state fermentation.
• It opens new avenues for applying hyperspectral imaging in the food and beverage industry.