Evaluation method for inter-chip uniformity in infrared mosaic arrays

Area of Science:

• Optics and Photonics
• Aerospace Engineering
• Image Processing

Background:

• Infrared mosaic arrays enhance detection resolution in aerospace.
• Evaluating inter-chip uniformity is crucial for optimal performance.
• Existing methods may lack accuracy or robustness.

Purpose of the Study:

• To propose a robust and accurate method for evaluating inter-chip uniformity in infrared mosaic arrays.
• To leverage human visual perception as a benchmark for uniformity assessment.
• To develop a model that guides stitching and non-uniformity correction.

Main Methods:

• Selected three key metrics: neighborhood difference (ND), Laplacian response (LR), and local standard deviation (LSTD).
• Fused these metrics using a support vector regression (SVR) model.
• Employed leave-one-out cross-validation (LOOCV) for generalization assessment.

Main Results:

• The proposed method demonstrated satisfactory generalization on unseen configurations.
• An ablation study confirmed the complementary nature and stable performance of the selected metrics.
• Validation on independent datasets confirmed the method's robustness.

Conclusions:

• The developed evaluation method accurately characterizes inter-chip uniformity in infrared mosaic arrays.
• This technique can guide the selection of stitching sequences and assess non-uniformity correction.
• The approach offers a reliable tool for improving infrared imaging systems.