Research on recognition of bedding system coverage rate using infrared thermal imaging

Area of Science:

• Thermal comfort research
• Sleep science
• Image processing

Background:

• Accurate assessment of bedding system coverage rate (BSCR) is crucial for understanding sleep thermal comfort.
• Existing methods for BSCR assessment are often contact-based and lack real-time capabilities.
• Developing a non-contact, real-time method is essential for objective sleep analysis.

Purpose of the Study:

• To develop and validate a non-contact method for real-time recognition of bedding system coverage rate (BSCR).
• To evaluate the influence of shooting angles and ambient temperatures on BSCR recognition accuracy.
• To correlate BSCR with subjective thermal perceptions and objective sleep quality.

Main Methods:

• Infrared thermal imaging combined with image segmentation algorithms (K-means clustering tested).
• Data collection from 30 subjects under varying shooting angles (45°, 60°, 90°) and ambient temperatures (23°C, 26°C, 29°C).
• Real-time BSCR recognition during summer, alongside collection of thermal sensation, thermal comfort votes, and sleep quality data.

Main Results:

• K-means clustering demonstrated high accuracy in BSCR recognition (MAE: 3.70, RMSE: 4.67, R²: 0.901).
• An optimal shooting angle of 60° was identified for the infrared camera.
• Ambient temperature within the tested range (23°C-29°C) did not significantly impact recognition accuracy.
• Average BSCR during comfortable sleep was 76.6%.
• Females exhibited a significantly higher BSCR (8.9%) compared to males (p < 0.01).

Conclusions:

• The proposed non-contact method effectively recognizes BSCR in real-time.
• The findings provide valuable data for developing non-contact sleep thermal comfort prediction models.
• Gender differences in BSCR highlight the need for personalized sleep environment assessments.