A Kalman-filtering approach to high dynamic range imaging for measurement applications

Area of Science:

• Computational photography
• Image processing
• Radiometric measurement

Background:

• High dynamic range imaging (HDRI) methods are crucial for capturing scenes exceeding single-exposure sensor limits.
• Existing HDRI techniques are often applied to radiance map construction in measurement applications.
• However, the design and evaluation of HDRI algorithms for these specific applications have been underexplored.

Purpose of the Study:

• To develop a novel high dynamic range imaging (HDRI) technique tailored for measurement applications.
• To introduce objective metrics for evaluating HDRI algorithm performance.
• To assess the effectiveness of the proposed HDRI method against established techniques.

Main Methods:

• Development of a new HDRI technique utilizing pixel-by-pixel Kalman filtering.
• Introduction of novel objective metrics for performance evaluation.
• Experimental validation of the proposed method using these metrics.

Main Results:

• The novel HDRI technique demonstrates significant improvements in signal-to-noise ratio, up to 9.4 dB.
• Radiometric accuracy is enhanced by as much as 29% compared to a classic method.
• The introduced objective metrics provide a robust framework for HDRI algorithm evaluation.

Conclusions:

• The proposed pixel-by-pixel Kalman filtering based HDRI technique offers superior performance for radiance map construction.
• This advancement addresses a critical gap in the evaluation of HDRI algorithms for measurement applications.
• The introduced metrics and method pave the way for more accurate and reliable high dynamic range imaging in scientific contexts.