Objective Simultaneous polarization imaging system can obtain real-time polarization images of moving targets, which has a wide range of applications. However, existing beam-splitting simultaneous polarization imaging systems have problems such as complex optical structures, high manufacturing and maintenance costs, low integration, and strict requirements for environmental stability, which greatly restrict their promotion and application in complex scenarios, such as small unmanned aerial vehicle remote sensing detection and industrial inspection. Therefore, it is crucial to develop a low-cost, miniaturized, and highly integrated simultaneous polarization imaging system to meet the demands of cost-sensitive applications.
Methods A three-camera simultaneous polarization imaging system with a parallel optical imaging architecture was designed. The system utilizes three miniature USB industrial camera modules and linear polarizers with fixed angles to construct a multi-channel optical framework. It utilizes 3D printing technology to deeply integrate polarization optical components, CMOS image sensors, an embedded Linux processing unit, and an MIPI touchscreen, achieving integration of polarization image acquisition and processing. To address system error sources, four calibration and correction methods were proposed: a multi-point correction method to resolve the non-consistency of pixel responses caused by fixed-pattern noise in camera modules; a linear response model-based method to correct radiation response non-consistency between channels; Malus' Law to calibrate polarizer angle errors; and the Scale-Invariant Feature Transform (SIFT) algorithm for image registration to correct geometric aberrations.
Results and Discussions After calibration and correction, the system performance was significantly optimized: the average polarization imaging non-uniformity was reduced to 0.94%; the average relative difference rate of radiation response between channels was lowered to −0.17%; polarizer angle calibration results showed an average error of 5.00° for the 45° polarizer and 0.00° for the 0° and 90° polarizers (Tab.2), and this error has been incorporated into Stokes parameter calculations for correction; the average image geometric aberration was reduced to 0.53 pixel (Tab.3). Meanwhile, comparison experiments with the prism beam-splitting simultaneous polarization imaging systems from Fluxdata (Fig.10-Fig.11, Tab.5) demonstrated that, while ensuring basic measurement accuracy, the proposed system achieves low-cost, miniaturized, and integrated multi-angle simultaneous polarization imaging, greatly enhancing its potential for promotion and application in cost-sensitive polarization imaging fields such as unmanned aerial vehicle remote sensing, industrial inspection, and education.
Conclusions To address the application limitations of traditional beam-splitting simultaneous polarization imaging systems, a three-camera simultaneous polarization imaging system with a parallel optical imaging architecture was successfully designed and implemented. By analyzing system error sources and implementing four calibration and correction methods, issues such as polarization imaging non-uniformity, inter-channel radiation response inconsistency, polarizer angle deviation, and geometric aberration were effectively suppressed. Experimental results showed that the system not only accurately captures target polarization information but also significantly reduces manufacturing costs and improves integration, environmental robustness, and modular maintainability compared to traditional systems. This system provides a feasible solution for the lightweight and popularization of multi-channel simultaneous polarization imaging technology, greatly enhancing its potential for application in cost-sensitive fields, including unmanned aerial vehicle remote sensing, industrial inspection, and education.
DownLoad: