Spectral Sensitivity Estimation with an Uncalibrated Diffraction Grating

2025-08-01

AI Summary

Camera spectral sensitivity calibration typically requires specialized narrowband filters or known-reflectance reference targets—resources often unavailable in practical settings. Method: This paper proposes a closed-form joint estimation method leveraging an uncalibrated, commercially available diffraction grating. By simultaneously capturing direct and diffracted light images, we formulate a physics-driven diffraction imaging model and jointly optimize the camera’s spectral response function along with critical grating parameters—including line density and tilt angle. Contribution/Results: To our knowledge, this is the first approach enabling analytical spectral sensitivity calibration using only an off-the-shelf diffraction grating, eliminating dependence on precision calibration equipment, standard reflectance targets, or prior spectral knowledge. Extensive evaluations on both synthetic and real-world data demonstrate significant performance gains over conventional reference-target-based methods, achieving high accuracy, strong robustness to noise and misalignment, and immediate engineering applicability.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Calibration & Uncertainty Quantification
• Search and Optimization: Distributed Search

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Security and Privacy: Large-scale security measurements
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

This paper introduces a practical and accurate calibration method for camera spectral sensitivity using a diffraction grating. Accurate calibration of camera spectral sensitivity is crucial for various computer vision tasks, including color correction, illumination estimation, and material analysis. Unlike existing approaches that require specialized narrow-band filters or reference targets with known spectral reflectances, our method only requires an uncalibrated diffraction grating sheet, readily available off-the-shelf. By capturing images of the direct illumination and its diffracted pattern through the grating sheet, our method estimates both the camera spectral sensitivity and the diffraction grating parameters in a closed-form manner. Experiments on synthetic and real-world data demonstrate that our method outperforms conventional reference target-based methods, underscoring its effectiveness and practicality.

Problem

• Estimates camera spectral sensitivity without calibration
• Uses uncalibrated diffraction grating for practical calibration
• Improves accuracy over traditional reference target methods

Innovation

• Uses uncalibrated diffraction grating sheet
• Estimates spectral sensitivity and grating parameters
• Closed-form method outperforms reference-based approaches