Forensic analysis of beverage stains using hyperspectral imaging.

Area of Science:

• Forensic Science
• Analytical Chemistry
• Spectroscopy

Background:

• Crime scene evidence analysis is crucial for forensic investigations.
• Beverage stains are common evidence, but their analysis can be challenging.
• Hyperspectral imaging (HSI) offers advanced capabilities for material analysis.

Purpose of the Study:

• To evaluate the potential of HSI for detecting and analyzing beverage stains on paper towels.
• To determine if HSI can identify common beverage types and predict stain age.
• To develop a rapid, non-contact, and non-invasive method for stain analysis.

Main Methods:

• Utilized 12 different beverages and four paper towel types to create sample stains.
• Employed HSI to capture spectral data from stains over 72 hours.
• Applied a convolutional auto-encoder for HSI data dimensionality reduction.
• Used a support vector machine (SVM) for classification on reduced data.
• Implemented volume-gradient-based band selection to identify relevant spectral bands.

Main Results:

• HSI effectively detected and analyzed beverage stains.
• Spectral changes over time allowed for prediction of stain age.
• Dimensionality reduction improved data processing and visualization.
• SVM classification achieved accurate stain identification.
• Volume-gradient-based band selection identified key spectral features.

Conclusions:

• HSI is a promising technique for rapid, non-contact, and non-invasive analysis of beverage stains in forensic investigations.
• The developed method can identify beverage types and estimate stain age.
• HSI provides valuable data for crime scene evidence analysis.