Radiometric imaging by double exposure and gain calibration

Area of Science:

• Optics
• Radiometry
• Image Processing

Background:

• Uncooled microbolometer arrays are cost-effective for radiometry but require frequent calibration due to temperature-dependent drifts.
• Standard gain calibration is feasible, but offset calibration is complex due to internal radiation loads.

Purpose of the Study:

• To develop a simplified computational optics approach for accurate temperature offset calibration in microbolometer-based thermal cameras.
• To eliminate the need for complicated calibration schemes by leveraging image blur variations.

Main Methods:

• A novel computational optics technique using two images with different known blur levels.
• Elimination of the object term from the image-formation equation to isolate sensor offset.
• Application of algebraic modeling for space-variant systems and solutions via direct inverse and iterative methods.

Main Results:

• Accurate restoration of radiometric values with errors as low as 1% (direct method) and 0.2% (iterative scheme).
• Demonstrated robustness against realistic lens positioning errors, with direct methods yielding an average error of 3.7%.

Conclusions:

• The proposed method significantly simplifies temperature offset calibration for uncooled thermal imaging systems.
• This approach enhances the reliability and accuracy of radiometric measurements from microbolometer arrays.