We introduce a ray mapping method for a binocular fringe projection system to fulfill a high-accurate measurement in the unknown refractive cross-media scenarios. This method builds the pixel-level ray mapping implicitly rather than launch the binocular calibration in pure air pinhole model and rectification for cross-media elaborately. We demonstrate how to use the orthogonal fringe patterns to fulfill the pixel-height mapping with the degraded epipolar constraint and parasitic reflections of multi-interface. Experimental results demonstrate that the proposed method enables high-precision 3D reconstruction in two typical cross-media scenarios, i.e., a glass box with and without water, with standard plane measurement errors consistently below 0.03 mm in all cases.

In optical three-dimensional (3D) measurement techniques, fringe projection is one of the most reliable techniques for recovering the shape of objects. One challenge of the fringe projection is the measurement of high dynamic range (HDR) surfaces. Current non-learning-based and learning-based methods typically require acquiring multiple phase-shifting (PS) fringe patterns to achieve high-precision HDR 3-D reconstruction, which leads to the limited measurement efficiency. To overcome this limitation, a single-shot HDR method (PI-SSM) based on color coding and spatial-frequency domain learning is proposed. There are two key contributions in this work. First, the physics-informed single-shot measurement framework which integrates hardware modulation and deep-learning algorithms in a novel way for HDR 3-D reconstruction is proposed. Specially, a color coding strategy is utilized to realize single-shot measurement, which acquires three fringe patterns simultaneously. Furthermore, by the network training, the color crosstalk problem caused by color coding can be suppressed, and the damaged phase of the HDR surfaces can be repaired. Second, to enhance fringe quality under severe HDR degradation, a spatial-frequency domain fringe enhancement network (SFENet) is specifically designed. SFENet restores degraded fringe by jointly modeling local noise-induced distortions in the spatial domain and enforcing global periodic consistency in the frequency domain. And a joint spatial-frequency loss further improves fringe enhancement quality and phase accuracy. Experiments demonstrate that the proposed PI-SSM method enables more accurate and efficient single-shot phase retrieval, exhibiting its excellent generalization to various unseen HDR surfaces.

Monocular, stable, low-computation-cost, and high-precision stereo imaging has been a key focus in the fields of robotics, autonomous driving, and machine vision. However, the majority of current stereo imaging research tends to prioritize the stereoscopic effects while neglecting the importance of computational efficiency. This significantly impacts the deployment and application of stereo imaging in industrial settings. In this paper, we propose a stereo imaging scheme based on depth from defocus, integrating the latest advancements in optical systems to overcome the computational efficiency limitations of existing algorithms. By introducing aberration analysis of the optical system, we re-derive the principles of depth from defocus to enhance both the accuracy and range of stereo imaging. Experimental results validate the effectiveness of the proposed method and the accuracy of the precision calculations. Finally, we analyze the impact of optical system parameters on the relevant methods. Theoretical derivations reveal the relationships between spatial resolution, depth resolution, and the range of stereo imaging in depth from defocus approaches, providing theoretical guidance for the design and deployment of such methods in practical application scenarios.

In three-dimensional (3D) shape measurement, codewords decoded from phase-coding patterns and wrapped phase extracted from sinusoidal patterns cannot maintain strict consistency at the edge of fringe period due to some unexpected influencing factors, such as defocus, noise and limited depth of field, resulting in occurrence of jump errors in absolute phase. To this end, a universal complementary strategy with high-frequency phase-coding patterns is proposed to avoid the jump errors for fringe projection profilometry (FPP). By splitting staggered fringe order sets from phase-coding patterns and dividing wrapped phase into different segment, the resulting fringe order can align with wrapped phase which guarantees the success of phase unwrapping. Furthermore, an improved method combining low-frequency phase-coding patterns with only one gray-coded fringe pattern is also used to elevate the accuracy rate of phase unwrapping. Simulation and experiments have shown that the proposed methods can greatly mitigate the phase jump errors, which implements in a simple and efficient way for 3D measurement of complex scenes.

Structured-illumination reflectance imaging (SIRI) has emerged as an effective approach for fruit quality assessment, particularly for detecting subsurface defects like bruises. However, conventional SIRI systems relying on 8-bit sinusoidal patterns face significant limitations in efficiency and robustness. Due to the bit-plane integration mechanism of digital light processing (DLP) projectors, the projection rate of 8-bit patterns is inherently restricted to a few 100 hertz, making them unsuitable for online quality inspection. Moreover, the demodulation accuracy is highly susceptible to undesired nonlinear gamma effects and strict hardware synchronization constraints. To overcome these limitations, this study presents an advanced effort for enhanced detection of apple bruises in SIRI, utilizing defocused 1-bit binary patterns to replace 8-bit sinusoidal patterns. Three classical binary patterns, including squared binary modulation (SBM), pulse-width-modulation (PWM), and error-diffusion dithering (EDD), were systematically evaluated under controlled projector defocus levels. Experimental results demonstrate that all three binary patterns effectively mitigate demodulation errors induced by nonlinear gamma effects and imprecise hardware synchronization. Among them, SBM yields superior accuracy at short fringe periods, whereas EDD performs best at long fringe periods. Overall, EDD stands out as the optimal choice, offering an alternative solution for online SIRI-based quality assessment of thin-skinned fruits. (c) 2026 Optica Publishing Group. All rights, including for text and data mining (TDM), Artificial Intelligence (AI) training, and similar technologies, are reserved.

In certain contexts, such as the conservation and restoration of relics, measurements must be carried out with the objects placed inside protective enclosures to ensure their preservation. This paper addresses the scenario in which an unknown refractive medium exists between the camera and the object under measurement and proposes a general-purpose calibration and reconstruction method. This method uses a ray model to directly determine height and lateral coordinates through phase information, achieving three-dimensional (3D) measurement without refraction errors. The method first establishes a mapping relationship between each pixel on the imaging plane and the emitted light rays then establishes a mapping between phase and height. Thus, when the phase is known, the height and lateral coordinates can be directly determined. To validate this method, we constructed a fringe projection profilometry (FPP) system using a glass box and a high-precision display calibration target. We performed 3D measurements on a standard plane and a regular sphere. The experimental results demonstrate the proposed technique's significant effectiveness and measurement accuracy.

Fourier single-pixel imaging (FSI), as a computational imaging method based on Fourier spectrum modulation, enables image reconstruction using only a single bucket detector. However, FSI faces a critical trade-off in practical applications: 8-bit grayscale projections incur excessive computational time, while 1-bit binary projections result in significantly degraded reconstruction quality. To address this, we propose a multi-bit error diffusion dithering algorithm that achieves intermediate bit depth quantization of Fourier bases within the 1-to-8-bit range. Furthermore, we develop a comprehensive evaluation framework to quantitatively assess the quality-speed trade-off, enabling identification of the optimal operating point under different sampling conditions. System simulations and experimental validation demonstrate that projection using 3-bit bases achieves reconstruction quality nearly identical to the 8-bit scheme (e.g., with a PSNR difference of <0.4 dB at 50% sampling rate), while offering a significant gain in projection efficiency. They also exhibit excellent adaptability across resolutions ranging from 322 to 2562 pixels. This research provides a novel, efficient encoding strategy and a systematic evaluation methodology for FSI, significantly advancing its application prospects in high-quality single-pixel imaging.

Ghost imaging (GI) is a typical computational imaging technique that reconstructs two-dimensional and three-dimensional images from one-dimensional bucket detector signals under structured light illumination. By utilizing single-pixel detection, this technology is particularly advantageous in low-light environments and in spectral regions (e.g., infrared, ultraviolet, or x-ray), where high-performance array detectors are often impractical or prohibitively expensive. However, traditional GI methods suffer from poor image reconstruction quality at low sampling rates and high hardware requirements; additionally, the generalization issues of data-driven deep learning methods limit their practical applications. Here, we propose a physics-driven Dual Untrained Ghost Imaging Neural Network (DUGIN). By integrating a "coarse-to-fine" dual-network architecture with the physical model, our method utilizes the deep image prior to achieve stable optimization and effectively escape local optima. Furthermore, a general affine scale correction module is designed to compensate for the intensity scale bias caused by normalization, further improving reconstruction fidelity. Simulation and experimental results demonstrate that DUGIN achieves high-fidelity natural image reconstruction at a 5% sampling rate, showing significantly reduced image noise and clearer details compared to traditional differential ghost imaging and recent physics-driven Ghost Imaging using Deep Neural Network Constraint methods. This study provides a novel framework for GI technology and paves the way for its practical application.

3-D shape measurement based on phase-shifting profilometry (PSP) has attracted extensive attention due to its high accuracy and noncontact characteristics. However, conventional PSP assumes a static scene during image acquisition, which severely restricts its applicability to dynamic scenarios. To overcome this limitation, this article proposes a spatial-consistent motion error compensation (SC-MEC) framework for dynamic 3-D imaging. Unlike existing methods that model motion-induced distortions solely as phase errors, the proposed approach jointly models pixel mismatch and phase error induced by object motion within a unified spatial framework. The motion vector of the target is estimated directly from reconstructed point clouds at adjacent time steps and, together with the calibrated spatial mapping between image and world coordinates, is used to initialize pixel displacement and phase deviation estimation. Based on this coupled error model, a global optimization strategy using simulated annealing (SA) is employed to iteratively refine both pixel- and phase-level compensation parameters by minimizing the geometric discrepancy between reconstructed point clouds. There are three major contributions in our method. First, a SC-MEC framework is proposed, which uses the spatial motion information of the target to jointly model the pixel mismatch and phase error and realizes the robust compensation under the complex nonuniform rigid-body motion. Second, the motion vectors and compensation parameters are estimated directly from the adjacent reconstructed point clouds, which need not be labeled and are hardware-independent, thereby providing wide applicability. Third, a new optimization paradigm is introduced to solve the coupling error model in structured-light measurement, and the SA algorithm is used to explore the global solution space to more accurately calculate the motion error in structured-light measurement.

Due to the light interference problem, three-dimensional (3D) reconstruction of complex scenarios remains challenging, especially when multiple reflections or light scattering exist. Although some attempts have been made to tackle this issue, there is still a lack of effective method simultaneously achieving high-accuracy and high-resolution reconstruction. In this work, we present a temporal-spatial modulation (TSM) framework to address those challenging issues by modulating illumination in both temporal and spatial dimensions. In the temporal dimension, TSM separates aliased light through frequency encoding and discrete Fourier analysis. In the spatial dimension, TSM leverages the physical imaging model and the Fourier power spectrum coefficients of direct light to reconstruct high-resolution fringe patterns and phase. Notably, the proposed method can be realized with off-the-shelf components such as the camera and projector, which is compatible with traditional structured light system. We validate our method in several challenging scenarios, including a sculpture with a mirror on its side, an L-shaped workpiece, and a ceramic bowl. The experiment results demonstrate the versatility of the proposed method.

In optical 3D sensing, high-efficiency and high-precision reconstruction of high-dynamic-range (HDR) scenes is highly crucial. Current HDR methods require either fusing images under multiple exposures or training network on massive datasets. To this end, a single-exposure full-resolution multi-polarization fringe enhancement method, establishing an effective hardware-algorithm collaboration mechanism for HDR 3D measurement for the first time, is proposed. Specifically, in the hardware part, a full-resolution multi-polarization (FM) imaging method is proposed. Unlike conventional multi-polarization modulation methods, the method avoids the 3/4-pixel loss and the instantaneous field-of-view error. Furthermore, by eliminating the specular reflection, more reliable scene information is captured, especially in overexposed areas, which significantly reduces the burden of fringe repair algorithm. In the algorithm part, a hybrid repair algorithm is proposed based on physics-informed zero-shot learning. The algorithm incorporates the physics priors of phase retrieval, FM fringe noise, and deep neural network into a dual-stream network architecture. This enables it to address underexposure and overexposure issues in the FM fringe without network pretraining on any dataset. Thus, the advantages of hardware modulation and algorithm optimization are fully integrated. Experiments on static and dynamic scenes with complex reflectivity demonstrate the superiority of the method in efficiency, resolution, fidelity, and generalization ability.

Event cameras, with their significantly higher dynamic range and sensitivity to intensity variations compared to frame cameras, provide new possibilities for 3D reconstruction in high-dynamic-range (HDR) scenes. However, the binary event data stream produced by event cameras presents significant challenges for achieving high-precision and efficient 3D reconstruction. In addressing these issues, we observe that the binary projection inherent to Gray-code-based 3D reconstruction naturally aligns with the event camera's imaging mechanism. However, achieving high-accuracy 3D reconstruction using a Gray code remains hindered by two key factors: inaccurate boundary extraction and the degradation of high-order dense Gray code patterns due to spatial blurring. For the first challenge, we propose an inverted Gray code strategy to improve region segmentation and recognition, achieving more precise and easily identifiable Gray code boundaries. For the second challenge, we introduce a spatial-shifting Gray-code encoding method. By spatially shifting Gray code patterns with lower encoding density, a combined encoding is achieved, enhancing the depth resolution and measurement accuracy. Experimental validation across general and HDR scenes demonstrates the effectiveness of the proposed methods.

Phase-shifting profilometry (PSP) has been widely used with its noncontact and high-precision characteristics in 3-D shape measurement. Motion of the measured object or assembly deviation of the mechanical projection device will, however, introduce phase-shift deviations in sequential phase shift images, which leads to phase errors and reduces 3-D measurement accuracy. This article proposes a generalized phase shift deviation estimation method for accurate 3-D shape measurement in phase-shifting profilometry. The N-step phase-shifting image sequences are grouped sequentially based on the image reusage strategy to efficiently calculate phases. The phase shift deviations are extracted from the difference information between adjacent phases without additional projection fringe, thus accurately calculating phase value. Experimental results demonstrate that the proposed method is generally applicable to the N-step PSP, which can effectively eliminate the phase error caused by the object motion and/or mechanical projector assembly errors and significantly improve the accuracy and efficiency of 3-D shape measurement.

Photoacoustic microscopy (PAM) offers high-resolution, non-invasive, and label-free imaging, making it invaluable for biomedical research. However, slow data acquisition and high sampling requirements remain key challenges that limit its broader applicability and scalability. We propose an Information-Efficient Photoacoustic Microscopy (IE-PAM) that jointly integrates sparse scanning encoding with neural network decoding to achieve high-quality reconstruction from extremely limited measurements. Specifically, IE-PAM employs a sparse-scanning acquisition scheme guided by random binary masks and reconstructs high-fidelity images using AFDU-Net, a custom-designed neural decoder trained on fully sampled ground truth data. Our system can faithfully recover detailed anatomical structures from as little as 1.5 % of the full sampling rate, corresponding to more than a 66-fold increase in acquisition efficiency. In in-vivo experiments on mouse ear vasculature, IE-PAM outperforms both traditional and learning-based baselines in fine vascular fidelity, artifact suppression, and robustness across varying sampling rates. By minimizing information redundancy at the acquisition stage and enabling accurate reconstruction from minimal data, IE-PAM provides a foundation for efficient, fast and scalable photoacoustic imaging in both preclinical and research applications.

In three-dimensional (3D) measurement, the motion of objects inevitably introduces errors, posing significant challenges to high-precision 3D reconstruction. Most existing algorithms for compensating motion-induced phase errors are tailored for object motion along the camera’s principal axis (Z direction), limiting their applicability in real-world scenarios where objects often experience complex combined motions in the X/Y and Z directions. To address these challenges, we propose a universal motion error compensation algorithm that effectively corrects both pixel mismatch and phase-shift errors, ensuring accurate 3D measurements under dynamic conditions. The method involves two key steps: first, pixel mismatch errors in the camera subsystem are corrected using adjacent coarse 3D point cloud data, aligning the captured data with the actual spatial geometry. Subsequently, motion-induced phase errors, observed as sinusoidal waveforms with a frequency twice that of the projection fringe pattern, are eliminated by applying the Hilbert transform to shift the fringes by π/2. Unlike conventional approaches that address these errors separately, our method provides a systematic solution by simultaneously compensating for camera-pixel mismatch and phase-shift errors within the 3D coordinate space. This integrated approach enhances the reliability and precision of 3D reconstruction, particularly in scenarios with dynamic and multidirectional object motions. The algorithm has been experimentally validated, demonstrating its robustness and broad applicability in fields such as industrial inspection, biomedical imaging, and real-time robotics. By addressing longstanding challenges in dynamic 3D measurement, our method represents a significant advancement in achieving high-accuracy reconstructions under complex motion environments.

The issue of scattering effect is common in imaging and optical 3D measurements, which introduces global illumination into the classical geometrical optics model. Separating the interested information from complex global-direct illumination often poses significant challenges. However, to suppress the global illumination, the conventional polarized modulated method generally requires manual adjustment to obtain multiple signals, which strictly limits the real-time detection and adaptability. To solve this problem, based on our analysis for the intensity distribution of four channels of the polarization camera from Malus's law, we establish a polarization angle shifting (PAS) model and further propose a separation strategy to efficiently achieve global-direct light transmission component pixel-by-pixel separation with a single exposure. And the proposed method eliminates the limitation of the linear polarization of the light source. With the proposed method, we could obtain contamination-free phase of scattering interface and global information imaging. Experimental results in a structured light measurement system confirm the effectiveness of the method for separation imaging, and the final de-scattering 3D phase results are also shown.

Among various 3D sensing technologies, multifocal structured light systems have gained significant attention for their ability of high-speed and large-depth-range measurement. However, as a non-standard structured light system, currently conventional calibration methods have model complexity or phase artifacts problems. To address these limitations, we propose a flexible pixel-wise calibration method for multifocal structured light system using LCD screen, which greatly improved measurement accuracy in a flexible way. Specifically, the phase artifacts problem of conventional calibration method is due to the use of high-contrast calibration board. To avoid this problem, we introduce an LCD screen to freely switch contrast in different calibration stage and initially establish pixel-wise phase-to-3D mapping model. Furthermore, considering the uneven surface of the LCD screen and camera calibration error, a standard plane along with plane fitting are used to refine phase-to-3D mapping model. Comparative experimental results show that the proposed method has obvious advantages over other methods.

In industrial 3-D metrology, the low signal-to-noise ratio (SNR) issue is commonly encountered, due to inappropriate illumination intensity, limited imaging dynamic range, or complex scene material, etc. Compared with nonlearning-based methods, deep-learning-based methods excel in efficiency and fidelity for the low SNR issue. However, most of them are data-driven, thus have limited generalization ability. Besides, they require advanced computing hardware for network training, greatly increasing the metrology cost. To tackle these problems, a physics-informed zero-shot learning (PZL) method with an ultralightweight neural network (UNN) is proposed for low-SNR scene measurement. There are two major contributions in our method. First, by blending physics priors for phase retrieval and fringe noise, a generalized PZL framework with a noisy-sinusoidal-component-to-noisy-sinusoidal-component (NS2NS) mapping is established. The low SNR issue of various challenging scenes including the low-illumination, high-dynamic-range, strong-ambient-light, and large-depth-range scenes is unified in a single enhancement framework. Moreover, no training dataset is required other than the degraded fringe itself, and the generalization ability for fringe enhancement is significantly improved. Second, based on the PZL framework, a symmetrized optimization strategy along with the UNN is proposed. Valid 3-D reconstruction of fine surface details can be achieved on computing-resource-constrained platforms, even on a CPU. Experiments verify the superiority of our method in efficiency, fidelity, generalization ability, and computing hardware cost. And to our knowledge, it is the first time such a simultaneous achievement has been accomplished.

High dynamic range (HDR) 3D measurement is a meaningful but challenging problem. Recently, many deep-learning-based methods have been proposed for the HDR problem. However, due to learning redundant fringe intensity information, their networks are difficult to converge for data with complex surface reflectivity and various illumination conditions, resulting in non-robust performance. To address this problem, we propose a physics-based supervised learning method. By introducing the physical model for phase retrieval, we design a novel, to the best of our knowledge, sinusoidal-component-to-sinusoidal-component mapping paradigm. Consequently, the scale difference of fringe intensity in various illumination scenarios can be eliminated. Compared with conventional supervised-learning methods, our method can greatly promote the convergence of the network and the generalization ability, while compared with the recently proposed unsupervised-learning method, our method can recover complex surfaces with much more details. To better evaluate our method, we specially design the experiment by training the network merely using the metal objects and testing the performance using different diffuse sculptures, metal surfaces, and their hybrid scenes. Experiments for all the testing scenarios have high-quality phase recovery with an STD error of about 0.03 rad, which reveals the superior generalization ability for complex reflectivity and various illumination conditions. Furthermore, the zoom-in 3D plots of the sculpture verify its fidelity on recovering fine details.