Unsupervised polarization imaging system polarization detection direction self-calibration method

CN122084544APending Publication Date: 2026-05-26BEIJING INST OF TECH
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
BEIJING INST OF TECH
Filing Date
2025-12-18
Publication Date
2026-05-26

AI Technical Summary

Technical Problem

Existing polarization imaging system calibration methods have stringent requirements. Supervised optimization algorithms require the true value of the polarization angle and are prone to calibration errors, while unsupervised optimization algorithms are only applicable to 4-channel systems and are difficult to achieve self-calibration of 3-channel and above systems when the true value of the polarization angle is unknown.

Method used

An unsupervised optimization algorithm is used to image scenes with unknown polarization information through a polarization imaging system, acquiring multiple sets of intensity images at different angles. Combining the mechanism of polarization degree stability, a gradient descent algorithm is used to adjust the polarization detection direction of the polarization system, and an error model is built to achieve self-calibration. This method is suitable for systems with 3 channels or more.

Benefits of technology

It achieves fast and accurate self-calibration of polarization system polarization direction under unknown true polarization angle conditions, with a wider range of applications, faster calculation speed, and stronger robustness, and is suitable for 3-channel and above systems.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122084544A_ABST
    Figure CN122084544A_ABST
Patent Text Reader

Abstract

The invention discloses an unsupervised polarization imaging system polarization detection direction self-calibration method, and belongs to the field of polarization imaging system calibration. The method comprises the following steps of: 1, imaging a fixed scene through a polarization imaging system to obtain scene light intensity, and calibrating a non-uniform coefficient k and a bias b of each channel through non-uniform correction to obtain normalized light intensity In; and 2, defining an analysis matrix A according to the Mueller matrix M of the polaroid, and calculating the polarization degree p according to the analysis matrix A and the normalized light intensity in the step 1. 3, taking the deviation angle as a to-be-optimized quantity, substituting the deviation angle into the step 2 for calculation, and writing p as a function of the sum I; an unsupervised optimization algorithm is adopted, under the condition that the true value of the polarization degree is unknown, p is calculated according to In by applying the mechanism of the stable polarization degree, the value is adjusted through a gradient descent algorithm, the root-mean-square error of p is made to be minimum, and self-calibration of the polarization detection direction of the polarization imaging system is achieved. The method has the advantages of being wide in application range, high in calibration speed, suitable for polarization systems with three or more channels and the like.
Need to check novelty before this filing date? Find Prior Art