Near-infrared monocular 3D computational polarization imaging of surfaces exhibiting nonuniform reflectance

Enhancing polarization 3D facial imaging: overcoming azimuth ambiguity without extra depth devices

Polarization 3D imaging has been a research hotspot in the field of 3D facial reconstruction because of its biosafety, high efficiency, and simplicity. However, the application of this technology is limited by the multi-valued problem of the azimuth angle of the normal vector. Currently, the most common method to overcome this limitation is to introduce additional depth techniques at the cost of reducing its applicability. This study presents a passive 3D polarization facial imaging method that does not require additional depth-capturing devices. It addresses the issue of azimuth ambiguity based on prior information about the target image's features. Specifically, by statistically analyzing the probability distribution of real azimuth angles, it is found that their quadrant distribution is closely related to the positions of facial feature points. Therefore, through facial feature detection, the polarized normal azimuth angle of each pixel can be accurately assigned to the corresponding quadrant, thus determining a precise unique normal vector and achieving accurate 3D facial reconstruction. Finally, our azimuth angle correction method was validated by simulated polarization imaging results, and it achieved accurate correction for over 75% of the global pixels without using additional depth techniques. Experimental results further indicate that this method can achieve polarization 3D facial imaging under natural conditions without extra depth devices, and the 3D results preserve edge details and texture information.

Multi-target distortion correction in 3D shape from polarization using a monocular camera system by deep neural networks

The shape from polarization is a noncontact 3D imaging method that shows great potential, but its application is limited by the monocular camera system and surface integration algorithm. This Letter proposes a novel, to the best of our knowledge, method that employs deep neural networks to enhance multi-target 3D reconstruction, making a significant advancement in the field. By constructing the relationship between targets' blur, distance, and clarity, the proposed method provides accurate spatial information while mitigating inaccuracies arising from the continuous model. Experiments show that the constructed neural network can help improve the multi-target 3D reconstruction quality compared with conventional methods.

Shape recovery from fusion of polarization binocular vision and shading

The shape from polarization can recover the fine texture of the target surface. However, the gradient field for shape recovery by polarization is ambiguous, which is caused by the multi-value of the azimuth angle. In response to the problem, a method of correcting the ambiguity by the fusion of polarization binocular vision and shading information is proposed in this paper. An iterative optimization algorithm is designed to estimate the direction of the light source, which provides the basis for the shading method to calculate the depth map. Additionally. the low-frequency depth map generated by binocular matching is used to correct the polarization gradient field. The polarization gradient field of the holes and small zenith angle regions in the binocular are corrected by the improved shading method. In the experiment, four different material target objects were used for shape recovery and compared with other methods. The results of the fusion method proposed are better in terms of fine texture. At the camera working distance of  700 mm, the resolving power performs well and demonstrates that changes in the depth of at least 0.1 mm can be distinguished from that recovery result.

Noise-aware infrared polarization image fusion based on salient prior with attention-guided filtering network

Infrared polarization image fusion integrates intensity and polarization information, producing a fused image that enhances visibility and captures crucial details. However, in complex environments, polarization imaging is susceptible to noise interference. Existing fusion methods typically use the infrared intensity (S0) and degree of linear polarization (DoLP) images for fusion but fail to consider the noise interference, leading to reduced performance. To cope with this problem, we propose a fusion method based on polarization salient prior, which extends DoLP by angle of polarization (AoP) and introduces polarization distance (PD) to obtain salient target features. Moreover, according to the distribution difference between S0 and DoLP features, we construct a fusion network based on attention-guided filtering, utilizing cross-attention to generate filter kernels for fusion. The quantitative and qualitative experimental results validate the effectiveness of our approach. Compared with other fusion methods, our method can effectively suppress noise interference and preserve salient target features.

Monocular polarized three-dimensional absolute depth reconstruction technology for multi-target scenes

The traditional polarization three-dimensional (3D) imaging technology has limited applications in the field of vision because it can only obtain the relative depth information of the target. Based on the principle of polarization stereo vision, this study combines camera calibration with a monocular ranging model to achieve high-precision recovery of the target's absolute depth information in multi-target scenes. Meanwhile, an adaptive camera intrinsic matrix prediction method is proposed to overcome changes in the camera intrinsic matrix caused by focusing on fuzzy targets outside the depth of field in multi-target scenes, thereby realizing monocular polarized 3D absolute depth reconstruction under dynamic focusing of targets at different depths. Experimental results indicate that the recovery error of monocular polarized 3D absolute depth information for the clear target is less than 10%, and the detail error is only 0.19 mm. Also, the precision of absolute depth reconstruction remains above 90% after dynamic focusing on the blurred target. The proposed monocular polarized 3D absolute depth reconstruction technology for multi-target scenes can broaden application scenarios of the polarization 3D imaging technology in the field of vision.

Analysis of Polarization Detector Performance Parameters on Polarization 3D Imaging Accuracy

Three-dimensional (3D) reconstruction of objects using the polarization properties of diffuse light on the object surface has become a crucial technique. Due to the unique mapping relation between the degree of polarization of diffuse light and the zenith angle of the surface normal vector, polarization 3D reconstruction based on diffuse reflection theoretically has high accuracy. However, in practice, the accuracy of polarization 3D reconstruction is limited by the performance parameters of the polarization detector. Improper selection of performance parameters can result in large errors in the normal vector. In this paper, the mathematical models that relate the polarization 3D reconstruction errors to the detector performance parameters including polarizer extinction ratio, polarizer installation error, full well capacity and analog-to-digital (A2D) bit depth are established. At the same time, polarization detector parameters suitable for polarization 3D reconstruction are provided by the simulation. The performance parameters we recommend include an extinction ratio ≥ 200, an installation error ∈ [-1°, 1°], a full-well capacity ≥ 100 Ke-, and an A2D bit depth ≥ 12 bits. The models provided in this paper are of great significance for improving the accuracy of polarization 3D reconstruction.

Estimation and removal of backscattered light with nonuniform polarization information in underwater environments

The nonuniform of polarization information of backscattered light has always been a neglected characteristic in polarization underwater imaging, but its accurate estimation plays an important role in the quality of imaging results. Traditional polarization imaging methods assume that the degree of polarization and angle of polarization of backscattered light are constant. In fact, the polarization information of backscattering light is gradual, this assumption makes traditional methods work only in a small area of the camera's field of view, in which the change of the polarization information of backscattered light can be ignored. In this paper, by analyzing the distribution of backscattered light, it is concluded that its polarization information has the characteristics of low-rank. Then, the degree of polarization and angle of polarization of backscattered light were estimated by low-rank and sparse matrix decomposition, and the clear scene was reconstructed. Experimental results show that the proposed method breaks through the limitation of the assumption of backscattered light in traditional polarization imaging method, and expands the detection field under the same conditions, which makes it possible to develop polarization underwater imaging method to the direction of large field of view detection.

Polarization-probe polarization-imaging system in near-infrared regime using a polarization grating

A polarization-probe polarization-imaging (PPPI) system was developed for the near-infrared (NIR) regime. This system comprises two components operating as a polarization generator and a polarization analyzer to enable polarization image capture under polarized light illumination. The captured polarization images contain considerable object information because the illuminating polarized light beams are affected by many of the Mueller matrix elements. By assembling the polarization camera using two liquid crystal retarders and a polarization grating, the PPPI system offers the potential to measure the Stokes parameters fully with a high extinction ratio, even in the NIR region. The PPPI system’s feasibility was demonstrated experimentally. Its dependence on the state of polarization (SoP) of the illuminating polarized light was discussed. The polarization image acquired by the PPPI system is strongly dependent on the illuminating light’s SoP, so the appropriate SoP must be selected for each object to enhance the polarization image contrast. This PPPI system should expand the range of polarization imaging applications, including LiDAR, product inspection, and bio-imaging.

Impact of color on polarization-based 3D imaging and countermeasures

Diffuse polarization-based 3D imaging has flourished with the ability to obtain the 3D shapes of objects without multiple detectors, active mode lighting, or complex mechanical structures, which are major drawbacks of other methods for 3D imaging in natural scenes. However, traditional polarization-based 3D imaging technology introduces color distortion when reconstructing the surface of multi-colored targets. We propose a polarization-based 3D imaging model to recover the 3D geometry of multi-colored Lambertian objects. In particular, chromaticity-based color removal theory is used to restore the intrinsic intensity, which is modulated only by the target shape, and we apply the recovered intrinsic intensity to address the orientation uncertainty of target normals due to azimuth ambiguity. Finally, we integrate the corrected normals to reconstruct high-precision 3D shapes. Experimental results demonstrate that the proposed model has the ability to reconstruct multi-colored Lambertian objects exhibiting non-uniform reflectance from single views under natural light conditions.

Enhancement of underwater vision by fully exploiting the polarization information from the Stokes vector

Underwater imaging method based on polarization information is extremely popular due to its ability to effectively remove the backscattered light. The Stokes vector contains the information of both the degree and angle of polarization of the light wave. However, this aspect has been rarely utilized in image reconstruction. In this study, an underwater polarimetric imaging model is established by fully exploiting this feature of Stokes vectors. The transmission of light wave is described in terms of the polarization information derived from the Stokes vector. Then, an optimization function is designed based on the independent characteristics of target light and backscattered light to estimate the target and backscattered field information. The real-world experiments and mean squared error analysis verify that the proposed method can remove the backscattered light and recover the target information accurately.