Yu Zhang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Yu Zhang has a total of 47 co-authors. Here are the top co-authors by number of shared publications:
- Jinlong Li
School of Physical Science and Technology, Southwest Jiaotong University, Chengdu 610031, China
9 shared publications - Lin Luo
Southwest Jiaotong University
9 shared publications - Jianping Peng
Southwest Jiaotong University
6 shared publications Lin Luo
6 shared publicationsXiaorong Gao
School of Physical Science and Technology, Southwest Jiaotong University, Chengdu 610031, China
5 shared publicationsJianping Peng
Department of Physical Science and Technology, Southwest Jiaotong University, Chengdu, China
4 shared publications
Latest Publications
- Open Access
- 18 November 2026
- Published by MDPI AG in Remote Sensing
- Vol. 18 (22) , 3966
Geometry-Aware Supervision for Connectivity-Preserving Road Extraction via SAM2
Road extraction from remote sensing imagery is a structured problem: a mask with high region overlap may still contain breaks that make the road network unusable, and the difficulty grows when annotated sources are small, heterogeneous, and unequally represented during joint training. We revisit this problem on a SAM2 encoder while keeping all added complexity in training rather than in the deployed network. First, we do not use SAM2 as a black box: its top-down pathway refines only the coarsest levels, leaving the quarter-resolution features our decoder needs semantically shallow, so a feature-enhancement neck restores the missing top-down path and a directional attention module matches the elongated geometry of roads. Second, we decompose road structure into five complementary geometric views—skeleton, direction, width, neighbourhood affinity, and endpoints—used as a shared training-time prior; each is a deterministic function of the segmentation mask, so the prior needs no extra annotation and no learned generator, and its heads are removed after training, leaving the inference graph free of auxiliary computation; the feature-enhancement neck and directional attention module remain in the deployed model, as they serve the segmentation head. A soft-clDice term supervises centrelines, and a square-root sampler rebalances the 5.2:1 update-frequency skew across sources. On Massachusetts, DeepGlobe, and CHN6-CUG, the model reaches an aggregate IoU of 0.7403 and CC-agreement of 0.6874, improving CC-agreement by 44% over D-LinkNet and 10% over the strongest same-backbone adaptation. We also make evaluation part of the contribution: we measure an empirical uncertainty scale for topology metrics and hold every claim to exceed it. Held to this standard, the prior’s CC-agreement gain clears the scale only in the scarce-annotation regime that limits operational mapping yet costs nothing to retain elsewhere. Finally, region overlap and CC-agreement rank models differently, so both metric families should be reported against a stated measurement scale.
- 18 November 2026
- Published by MDPI AG in Remote Sensing
- Vol. 18 (22) , 3966
Open AccessGeometry-Aware Supervision for Connectivity-Preserving Road Extraction via SAM2
Road extraction from remote sensing imagery is a structured problem: a mask with high region overlap may still contain breaks that make the road network unusable, and the difficulty grows when annotated sources are small, heterogeneous, and unequally represented during joint training. We revisit this problem on a SAM2 encoder while keeping all added complexity in training rather than in the deployed network. First, we do not use SAM2 as a black box: its top-down pathway refines only the coarsest levels, leaving the quarter-resolution features our decoder needs semantically shallow, so a feature-enhancement neck restores the missing top-down path and a directional attention module matches the elongated geometry of roads. Second, we decompose road structure into five complementary geometric views—skeleton, direction, width, neighbourhood affinity, and endpoints—used as a shared training-time prior; each is a deterministic function of the segmentation mask, so the prior needs no extra annotation and no learned generator, and its heads are removed after training, leaving the inference graph free of auxiliary computation; the feature-enhancement neck and directional attention module remain in the deployed model, as they serve the segmentation head. A soft-clDice term supervises centrelines, and a square-root sampler rebalances the 5.2:1 update-frequency skew across sources. On Massachusetts, DeepGlobe, and CHN6-CUG, the model reaches an aggregate IoU of 0.7403 and CC-agreement of 0.6874, improving CC-agreement by 44% over D-LinkNet and 10% over the strongest same-backbone adaptation. We also make evaluation part of the contribution: we measure an empirical uncertainty scale for topology metrics and hold every claim to exceed it. Held to this standard, the prior’s CC-agreement gain clears the scale only in the scarce-annotation regime that limits operational mapping yet costs nothing to retain elsewhere. Finally, region overlap and CC-agreement rank models differently, so both metric families should be reported against a stated measurement scale.
- Open Access
- 1 October 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Photonics Journal
- Vol. 18 (5) , 6700209
Single-Shot 3D Shape Measurement Using Speckle-Embedded Colored Fringe Patterns With Neighborhood-Assisted Diffusion Compensation
- 1 October 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Photonics Journal
- Vol. 18 (5) , 6700209
Open AccessSingle-Shot 3D Shape Measurement Using Speckle-Embedded Colored Fringe Patterns With Neighborhood-Assisted Diffusion Compensation
- Open Access
- 3 June 2026
- Published by Elsevier in SSRN Electronic Journal
Resolving Geometric and Radiometric Singularities in Photometric Stereo: An Optimized Dual-Spectral Framework for Industrial Surface Imaging
Dual-spectral line-scan photometric stereo is widely utilized for the high-speed inspection of planar industrial surfaces. However, conventional coplanar setups exhibit directional blindness, as their rank-deficient illumination geometry yields near-zero sensitivity to surface gradients along the motion axis. Furthermore, uncalibrated sensor non-linearity violates the linear assumptions of ratiometric solvers, introducing low-frequency undulating artifacts and amplifying stochastic noise. To resolve these coupled geometric and radiometric limitations, we propose an optical-algorithmic co-design framework. First, we theoretically demonstrate that modifying the illumination azimuth to an optimal diagonal configuration breaks the coplanar singularity, transitioning the system into a well-posed, mathematically isotropic state with a minimized condition number. Second, to restore the physical validity of the Lambertian baseline, we formulate a reflectance-equivalent radiometric linearization protocol that suppresses non-linear intensity-distortion crosstalk. Finally, to overcome shading cancellation under symmetric topographies without introducing photometric crosstalk, we propose an adaptive orthogonal-weighted fusion (AOWF) mechanism. By employing a variance-aware gating strategy driven by the differential photometric signal, AOWF dynamically decouples orthogonal geometric cues from sum-space background noise. Comprehensive evaluations across diverse industrial substrates validate that this synergistic framework extracts high-fidelity topological structures, demonstrating quantitative improvements in defect separability and edge sharpness compared to conventional methods.
- 3 June 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessResolving Geometric and Radiometric Singularities in Photometric Stereo: An Optimized Dual-Spectral Framework for Industrial Surface Imaging
Dual-spectral line-scan photometric stereo is widely utilized for the high-speed inspection of planar industrial surfaces. However, conventional coplanar setups exhibit directional blindness, as their rank-deficient illumination geometry yields near-zero sensitivity to surface gradients along the motion axis. Furthermore, uncalibrated sensor non-linearity violates the linear assumptions of ratiometric solvers, introducing low-frequency undulating artifacts and amplifying stochastic noise. To resolve these coupled geometric and radiometric limitations, we propose an optical-algorithmic co-design framework. First, we theoretically demonstrate that modifying the illumination azimuth to an optimal diagonal configuration breaks the coplanar singularity, transitioning the system into a well-posed, mathematically isotropic state with a minimized condition number. Second, to restore the physical validity of the Lambertian baseline, we formulate a reflectance-equivalent radiometric linearization protocol that suppresses non-linear intensity-distortion crosstalk. Finally, to overcome shading cancellation under symmetric topographies without introducing photometric crosstalk, we propose an adaptive orthogonal-weighted fusion (AOWF) mechanism. By employing a variance-aware gating strategy driven by the differential photometric signal, AOWF dynamically decouples orthogonal geometric cues from sum-space background noise. Comprehensive evaluations across diverse industrial substrates validate that this synergistic framework extracts high-fidelity topological structures, demonstrating quantitative improvements in defect separability and edge sharpness compared to conventional methods.
- Open Access
- 24 March 2026
- Published by Elsevier in SSRN Electronic Journal
Single-shot 3D shape measurement using speckle-embedded colored fringe patterns with neighborhood-assisted diffusion compensation
Single-shot structured-light techniques are essential for dynamic 3D measurement. However, speckle-assisted correspondence matching is often affected by correlation mismatches, which severely degrade disparity estimation and reconstruction accuracy. In this paper, a robust single-shot 3D measurement method based on color fringe-speckle pattern is proposed to address correlation-induced mismatches in speckle-assisted correspondence. To suppress unreliable correspondences, a region-based disparity optimization framework is introduced. Specifically, reliable disparity candidates are first identified via density-based clustering within continuous phase regions, effectively eliminating severe mismatches caused by false correlation peaks. Subsequently, a neighborhood-assisted diffusion compensation strategy is applied by exploiting local spatial consistency to recover unreliable or missing disparities, thereby improving disparity completeness. Experimental results on isolated and complex static objects demonstrate that the proposed method significantly improves reconstruction quality. Dynamic experiments involving depth-direction translation and rotational motion further verify that the proposed approach enables stable single-shot 3D reconstruction without noticeable motion-induced artifacts. Quantitative evaluation using a standard sphere shows that the proposed method achieves an RMS accuracy of 0.1 mm. The proposed method enhances the reliability of speckle-assisted single-shot structured-light measurement without requiring additional projected patterns, making it well suited for dynamic 3D sensing applications.
- 24 March 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessSingle-shot 3D shape measurement using speckle-embedded colored fringe patterns with neighborhood-assisted diffusion compensation
Single-shot structured-light techniques are essential for dynamic 3D measurement. However, speckle-assisted correspondence matching is often affected by correlation mismatches, which severely degrade disparity estimation and reconstruction accuracy. In this paper, a robust single-shot 3D measurement method based on color fringe-speckle pattern is proposed to address correlation-induced mismatches in speckle-assisted correspondence. To suppress unreliable correspondences, a region-based disparity optimization framework is introduced. Specifically, reliable disparity candidates are first identified via density-based clustering within continuous phase regions, effectively eliminating severe mismatches caused by false correlation peaks. Subsequently, a neighborhood-assisted diffusion compensation strategy is applied by exploiting local spatial consistency to recover unreliable or missing disparities, thereby improving disparity completeness. Experimental results on isolated and complex static objects demonstrate that the proposed method significantly improves reconstruction quality. Dynamic experiments involving depth-direction translation and rotational motion further verify that the proposed approach enables stable single-shot 3D reconstruction without noticeable motion-induced artifacts. Quantitative evaluation using a standard sphere shows that the proposed method achieves an RMS accuracy of 0.1 mm. The proposed method enhances the reliability of speckle-assisted single-shot structured-light measurement without requiring additional projected patterns, making it well suited for dynamic 3D sensing applications.
- 18 December 2025
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 37 (1) , 015406
Single-shot 3D measurement with global phase offset correction based on stereo color phase-shifting profilometry
Phase-shifting profilometry (PSP) plays a dominant role in the field of three-dimensional (3D) surface-shape measurement due to its high accessible measurement accuracy and spatial resolution. However, PSP relies on multiple projections, which makes it prone to motion-induced errors and thus its performance is limited in dynamic scenes. In this paper, we introduce a novel single-shot 3D measurement method based on stereo color PSP. This method requires no additional images, benefiting from geometric constraints and colored phase-shifted fringes. The phase errors caused by color crosstalk and gamma nonlinearity are corrected by the chord distribution equalization method and the cross-correlation-based global phase offset correction method. The point cloud obtained using the corrected phase is reprojected onto the second camera plane, and the corresponding matching points are found on the epipolar lines within a limited search range. After stereo correspondence, the final 3D point cloud of the measured object is reconstructed based on stereo technology. Three dynamic scenes (a changing palm, a rotating Beethoven statue on a platform, and a rotating electric fan) are assessed to verify the robustness and accuracy of the proposed technique in a range of dynamic and complex environments.
- 18 December 2025
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 37 (1) , 015406
Single-shot 3D measurement with global phase offset correction based on stereo color phase-shifting profilometry
Phase-shifting profilometry (PSP) plays a dominant role in the field of three-dimensional (3D) surface-shape measurement due to its high accessible measurement accuracy and spatial resolution. However, PSP relies on multiple projections, which makes it prone to motion-induced errors and thus its performance is limited in dynamic scenes. In this paper, we introduce a novel single-shot 3D measurement method based on stereo color PSP. This method requires no additional images, benefiting from geometric constraints and colored phase-shifted fringes. The phase errors caused by color crosstalk and gamma nonlinearity are corrected by the chord distribution equalization method and the cross-correlation-based global phase offset correction method. The point cloud obtained using the corrected phase is reprojected onto the second camera plane, and the corresponding matching points are found on the epipolar lines within a limited search range. After stereo correspondence, the final 3D point cloud of the measured object is reconstructed based on stereo technology. Three dynamic scenes (a changing palm, a rotating Beethoven statue on a platform, and a rotating electric fan) are assessed to verify the robustness and accuracy of the proposed technique in a range of dynamic and complex environments.
- Open Access
- 23 October 2024
- Published by Springer Nature in Scientific Reports
- Vol. 14 (1) , 1-14
Dual frequency composite pattern temporal phase unwrapping for 3D surface measurement
Phase shifting profilometry (PSP) is widely used in three-dimensional (3D) optical metrology applications such as mechanical engineering, industrial monitoring, computer vision, and biomedicine because of its high measurement accuracy. In PSP, multi-frequency temporal phase unwrapping (TPU) plays a dominant role due to its high-accuracy measurement of surfaces with discontinuities and isolated objects. However, it requires a large number of fringe patterns. To reduce the number of required patterns, a new dual-frequency composite-pattern TPU method is developed, which needs only three patterns. The new method combines two fringe patterns of different frequencies into three composite fringe patterns, reconstructs rough 3D points by demodulating the low-frequency fringe pattern using geometric constraints, and reprojects it to obtain a rough low-frequency absolute phase, which is then refined to correct edge errors. Finally, the high-frequency wrapped phase is unwrapped under the guidance of the refined low-frequency phase, thereby the accurate 3D points are reconstructed based on stereo-vision technology. Experimental results demonstrated that the proposed method can achieve 3D surface measurement with only three images, which greatly improves the measurement efficiency. The proposed method has great application potential for the rapid 3D measurement of discontinuous surfaces and isolated objects.
- 23 October 2024
- Published by Springer Nature in Scientific Reports
- Vol. 14 (1) , 1-14
Open AccessDual frequency composite pattern temporal phase unwrapping for 3D surface measurement
Phase shifting profilometry (PSP) is widely used in three-dimensional (3D) optical metrology applications such as mechanical engineering, industrial monitoring, computer vision, and biomedicine because of its high measurement accuracy. In PSP, multi-frequency temporal phase unwrapping (TPU) plays a dominant role due to its high-accuracy measurement of surfaces with discontinuities and isolated objects. However, it requires a large number of fringe patterns. To reduce the number of required patterns, a new dual-frequency composite-pattern TPU method is developed, which needs only three patterns. The new method combines two fringe patterns of different frequencies into three composite fringe patterns, reconstructs rough 3D points by demodulating the low-frequency fringe pattern using geometric constraints, and reprojects it to obtain a rough low-frequency absolute phase, which is then refined to correct edge errors. Finally, the high-frequency wrapped phase is unwrapped under the guidance of the refined low-frequency phase, thereby the accurate 3D points are reconstructed based on stereo-vision technology. Experimental results demonstrated that the proposed method can achieve 3D surface measurement with only three images, which greatly improves the measurement efficiency. The proposed method has great application potential for the rapid 3D measurement of discontinuous surfaces and isolated objects.
- 31 July 2024
- Published by IOP Publishing in Journal of Optics
- Vol. 26 (9) , 095704
Online nonlinearity elimination for fringe projection profilometry using slope intensity coding
The nonlinearity effect in the system of fringe projection profilometry can cause the non-sinusoidal deviation of the fringe patterns, inducing ripple-like phase errors and further affecting measurement accuracy. This paper presents an online nonlinearity elimination method based on slope intensity coding. Two sequences of sinusoidal phase-shifting fringe patterns with different frequencies, and one slope intensity pattern with one uniform intensity pattern are projected. The equations for the nonlinearity response are established using the defined mean and modulation parameters, the captured uniform intensity and two extracted background intensities. The nonlinearity response coefficients determined by solving the equations are used for pixel-wise nonlinearity correction on the captured images, which are employed for computing the wrapped phase, and further obtaining continuous phase by the multi-frequency phase unwrapping method. Experimental results demonstrate that the proposed method can eliminate the nonlinearity-induced phase error online by using fewer images and maintain the reliability of phase unwrapping in the measurement of isolated objects with complex surfaces.
- 31 July 2024
- Published by IOP Publishing in Journal of Optics
- Vol. 26 (9) , 095704
Online nonlinearity elimination for fringe projection profilometry using slope intensity coding
The nonlinearity effect in the system of fringe projection profilometry can cause the non-sinusoidal deviation of the fringe patterns, inducing ripple-like phase errors and further affecting measurement accuracy. This paper presents an online nonlinearity elimination method based on slope intensity coding. Two sequences of sinusoidal phase-shifting fringe patterns with different frequencies, and one slope intensity pattern with one uniform intensity pattern are projected. The equations for the nonlinearity response are established using the defined mean and modulation parameters, the captured uniform intensity and two extracted background intensities. The nonlinearity response coefficients determined by solving the equations are used for pixel-wise nonlinearity correction on the captured images, which are employed for computing the wrapped phase, and further obtaining continuous phase by the multi-frequency phase unwrapping method. Experimental results demonstrate that the proposed method can eliminate the nonlinearity-induced phase error online by using fewer images and maintain the reliability of phase unwrapping in the measurement of isolated objects with complex surfaces.
- Open Access
FR-PatchCore: An Industrial Anomaly Detection Method for Improving Generalization
In recent years, a multitude of self-supervised anomaly detection algorithms have been proposed. Among them, PatchCore has emerged as one of the state-of-the-art methods on the widely used MVTec AD benchmark due to its efficient detection capabilities and cost-saving advantages in terms of labeled data. However, we have identified that the PatchCore similarity principal approach faces significant limitations in accurately locating anomalies when there are positional relationships between similar samples, such as rotation, flipping, or misaligned pixels. In real-world industrial scenarios, it is common for samples of the same class to be found in different positions. To address this challenge comprehensively, we introduce Feature-Level Registration PatchCore (FR-PatchCore), which serves as an extension of the PatchCore method. FR-PatchCore constructs a feature matrix that is extracted into the memory bank and continually updated using the optimal negative cosine similarity loss. Extensive evaluations conducted on the MVTec AD benchmark demonstrate that FR-PatchCore achieves an impressive image-level anomaly detection AUROC score of up to 98.81%. Additionally, we propose a novel method for computing the mask threshold that enables the model to scientifically determine the optimal threshold and accurately partition anomalous masks. Our results highlight not only the high generalizability but also substantial potential for industrial anomaly detection offered by FR-PatchCore.
Open AccessFR-PatchCore: An Industrial Anomaly Detection Method for Improving Generalization
In recent years, a multitude of self-supervised anomaly detection algorithms have been proposed. Among them, PatchCore has emerged as one of the state-of-the-art methods on the widely used MVTec AD benchmark due to its efficient detection capabilities and cost-saving advantages in terms of labeled data. However, we have identified that the PatchCore similarity principal approach faces significant limitations in accurately locating anomalies when there are positional relationships between similar samples, such as rotation, flipping, or misaligned pixels. In real-world industrial scenarios, it is common for samples of the same class to be found in different positions. To address this challenge comprehensively, we introduce Feature-Level Registration PatchCore (FR-PatchCore), which serves as an extension of the PatchCore method. FR-PatchCore constructs a feature matrix that is extracted into the memory bank and continually updated using the optimal negative cosine similarity loss. Extensive evaluations conducted on the MVTec AD benchmark demonstrate that FR-PatchCore achieves an impressive image-level anomaly detection AUROC score of up to 98.81%. Additionally, we propose a novel method for computing the mask threshold that enables the model to scientifically determine the optimal threshold and accurately partition anomalous masks. Our results highlight not only the high generalizability but also substantial potential for industrial anomaly detection offered by FR-PatchCore.
- Open Access
- 20 November 2023
- Published by MDPI AG in Applied Sciences
- Vol. 13 (22) , 12513
WIG-Net: Wavelet-Based Defocus Deblurring with IFA and GCN
Although the existing deblurring methods for defocused images are capable of approximately recovering clear images, they still exhibit certain limitations, such as ringing artifacts and remaining blur. Along these lines, in this work, a novel deep-learning-based method for image defocus deblurring was proposed, which can be applied to medical images, traffic monitoring, and other fields. The developed approach is equipped with wavelet transform, an iterative filter adaptive module, and graph neural network and was specifically designed for handling defocused blur. Our network exhibits excellent properties in preserving the original information during the restoration of clear images, thereby enhancing its capability to spatially address varying blurriness and improving the quality of deblurring. From the acquired experimental results, the superiority of the introduced method in the context of image defocus deblurring compared to the majority of the existing algorithms was clearly demonstrated.
- 20 November 2023
- Published by MDPI AG in Applied Sciences
- Vol. 13 (22) , 12513
Open AccessWIG-Net: Wavelet-Based Defocus Deblurring with IFA and GCN
Although the existing deblurring methods for defocused images are capable of approximately recovering clear images, they still exhibit certain limitations, such as ringing artifacts and remaining blur. Along these lines, in this work, a novel deep-learning-based method for image defocus deblurring was proposed, which can be applied to medical images, traffic monitoring, and other fields. The developed approach is equipped with wavelet transform, an iterative filter adaptive module, and graph neural network and was specifically designed for handling defocused blur. Our network exhibits excellent properties in preserving the original information during the restoration of clear images, thereby enhancing its capability to spatially address varying blurriness and improving the quality of deblurring. From the acquired experimental results, the superiority of the introduced method in the context of image defocus deblurring compared to the majority of the existing algorithms was clearly demonstrated.
- 13 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 26-30
A Domain-adaptive Method of Binocular Disparity Estimation for Bottom Components of the Train
Disparity estimation of binocular images is an important branch of depth reconstruction. The neural network method based on depth learning shows excellent performance in binocular stereo vision, which is far beyond the traditional binocular depth estimation algorithm. Most binocular stereo matching networks rely on training the dataset of the target scene, and can not predict the disparity of bottom components of the train directly. For this problem, we propose a domain-adaptive network, which includes two types of feature extraction encoding layers, a cost volume module that fuses a pair of image features, and a disparity prediction module based on multi-scale convolution layer. Moreover, based on the results of network prediction, we use a consistency checking method to further refine the disparity and enhance the reliability of the estimation results. Through experiments in different scenes, our method shows more accuracy and generalization than ordinary stereo matching networks, and has good effect of stereo visualization with less memory consumption in real train scenes. At the same time, our optimization method of consistency reduces the error of prediction and improves the confidence of disparity estimation for train bottom as a whole.
- 13 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 26-30
A Domain-adaptive Method of Binocular Disparity Estimation for Bottom Components of the Train
Disparity estimation of binocular images is an important branch of depth reconstruction. The neural network method based on depth learning shows excellent performance in binocular stereo vision, which is far beyond the traditional binocular depth estimation algorithm. Most binocular stereo matching networks rely on training the dataset of the target scene, and can not predict the disparity of bottom components of the train directly. For this problem, we propose a domain-adaptive network, which includes two types of feature extraction encoding layers, a cost volume module that fuses a pair of image features, and a disparity prediction module based on multi-scale convolution layer. Moreover, based on the results of network prediction, we use a consistency checking method to further refine the disparity and enhance the reliability of the estimation results. Through experiments in different scenes, our method shows more accuracy and generalization than ordinary stereo matching networks, and has good effect of stereo visualization with less memory consumption in real train scenes. At the same time, our optimization method of consistency reduces the error of prediction and improves the confidence of disparity estimation for train bottom as a whole.