A DNN-Based UVI Calculation Method Using Representative Color Information of Sun Object Images.

Area of Science:

• Environmental Science
• Atmospheric Science
• Computer Science

Background:

• Increasing interest in environmental factors like ultraviolet (UV) radiation for health.
• UV radiation is crucial but difficult to measure directly due to its invisible wavelengths.
• Existing UV measurement devices are costly and inconvenient for general users.

Purpose of the Study:

• To propose a deep neural network (DNN)-based method for calculating the ultraviolet index (UVI).
• To enable accessible measurement of ambient UV radiation using sun object images.
• To develop a cost-effective and convenient UVI calculation approach.

Main Methods:

• Utilized Mask-region-based convolutional neural networks (R-CNN) to extract sun object regions from sky images.
• Detected representative RGB color values from the extracted sun object regions.
• Constructed a DNN model to calculate UVI using RGB values, sun's altitude angle, and azimuth.

Main Results:

• The proposed DNN-based method accurately calculates UVI from sun object image colors.
• The method achieved a mean absolute error of 0.3 during testing in spring and autumn.
• Demonstrated the feasibility of estimating UV radiation intensity indirectly.

Conclusions:

• The developed DNN model provides an accurate and accessible way to determine UVI.
• This method overcomes the limitations of direct UV measurement devices.
• Facilitates better public awareness and management of UV exposure for health benefits.