Qing He
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Qing He has a total of 35 co-authors. Here are the top co-authors by number of shared publications:
- Hongji Meng
Institute of Automation Instrument, College of Information Science and Engineering, Northeastern University, Shenyang, China
4 shared publications - Zhi Xie
College of Information Science and Engineering, Northeastern University, Shenyang, China
4 shared publications - Jian Yang
School of Information Science and Engineering, Northeastern University, Shenyang 110819, China
2 shared publications Qingying Li
Northeastern Univ, Coll Informat Sci & Engn, Shenyang, Peoples R China
2 shared publicationsXu Liu
Northeastern Univ, Coll Informat Sci & Engn, Shenyang, Peoples R China
2 shared publications- Zhiwen Ji
College of Information Science and Engineering, Northeastern University, Shenyang, China
2 shared publications
Latest Publications
- 13 November 2026
- Published by Springer Nature in Journal of Intelligent Manufacturing
- p. 1-20
An edge-based online mandrel surface defect inspection system under strong industrial interference
- 13 November 2026
- Published by Springer Nature in Journal of Intelligent Manufacturing
- p. 1-20
An edge-based online mandrel surface defect inspection system under strong industrial interference
- Open Access
- 17 February 2026
- Published by Elsevier in SSRN Electronic Journal
Real-Time Crack Segmentation and Geometric Measurement of Mandrel Bars Based on an Improved YOLO Framework
Surface cracks on mandrel bars seriously affect product quality and production stability in seamless steel pipe manufacturing. To address the limitations of bounding-box-based inspection methods in providing accurate geometric characterization, this study proposes a lightweight, segmentation-based crack inspection framework for real-time industrial applications. By reformulating crack inspection from detection to pixel-level segmentation, the proposed approach enables precise localization and quantitative measurement of crack geometry. A dedicated industrial dataset and a background-oriented sample-balancing strategy are adopted to enhance robustness under background-dominated inspection conditions. Experimental results show that the proposed method achieves a segmentation performance of 90.5% mAP@0.5, with a low false negative rate of 1.73%, while maintaining real-time inference at 195 FPS with only 3.01 GFLOPs. Deployment on an actual production line further demonstrates the reliability and engineering value of the proposed framework for on-line crack inspection and geometric evaluation of mandrel bars.
- 17 February 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessReal-Time Crack Segmentation and Geometric Measurement of Mandrel Bars Based on an Improved YOLO Framework
Surface cracks on mandrel bars seriously affect product quality and production stability in seamless steel pipe manufacturing. To address the limitations of bounding-box-based inspection methods in providing accurate geometric characterization, this study proposes a lightweight, segmentation-based crack inspection framework for real-time industrial applications. By reformulating crack inspection from detection to pixel-level segmentation, the proposed approach enables precise localization and quantitative measurement of crack geometry. A dedicated industrial dataset and a background-oriented sample-balancing strategy are adopted to enhance robustness under background-dominated inspection conditions. Experimental results show that the proposed method achieves a segmentation performance of 90.5% mAP@0.5, with a low false negative rate of 1.73%, while maintaining real-time inference at 195 FPS with only 3.01 GFLOPs. Deployment on an actual production line further demonstrates the reliability and engineering value of the proposed framework for on-line crack inspection and geometric evaluation of mandrel bars.
- 12 June 2025
- Published by Elsevier in Precision Engineering
- Vol. 96, 170-181
Direct phase aberration compensation through gradient fitting for wafer surface 3D measurement using digital holography
- 12 June 2025
- Published by Elsevier in Precision Engineering
- Vol. 96, 170-181
Direct phase aberration compensation through gradient fitting for wafer surface 3D measurement using digital holography
- 4 December 2024
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 36 (1) , 015218
A 3D calibration method for binocular vision system
Binocular vision system calibration is extremely important in three-dimensional (3D) measurement. During the calibration process, the establishment of an appropriate objective function has a direct impact on the calibration accuracy. For common two-dimensional (2D) calibration methods, differences in 2D calibration and 3D measurement coordinate systems result in loss of accuracy. Furthermore, for the newly proposed 3D calibration methods, the epipolar constraints in the 2D pixel coordinate system and the reconstruction errors in the 3D physical space need to be balanced by weights, resulting in uncertainties in the calibration results. To address these issues, we propose a novel calibration method that minimizes the distances between the control point and the left and right camera rays. The distances between the control points and the camera rays describe the 3D reconstruction errors, and the distances between the left camera and the right camera rays describe the geometry information between the two cameras to ensure that the epipolar constraints are satisfied. As a result, these two types of constraints/errors are measured under the same 3D coordinate system. Therefore, the epipolar constraints and the reconstruction errors are naturally balanced. We have conducted computer simulations and real experiments, and the results show that the proposed calibration method is effective in improving the accuracy compared with the state-of-the-art methods.
- 4 December 2024
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 36 (1) , 015218
A 3D calibration method for binocular vision system
Binocular vision system calibration is extremely important in three-dimensional (3D) measurement. During the calibration process, the establishment of an appropriate objective function has a direct impact on the calibration accuracy. For common two-dimensional (2D) calibration methods, differences in 2D calibration and 3D measurement coordinate systems result in loss of accuracy. Furthermore, for the newly proposed 3D calibration methods, the epipolar constraints in the 2D pixel coordinate system and the reconstruction errors in the 3D physical space need to be balanced by weights, resulting in uncertainties in the calibration results. To address these issues, we propose a novel calibration method that minimizes the distances between the control point and the left and right camera rays. The distances between the control points and the camera rays describe the 3D reconstruction errors, and the distances between the left camera and the right camera rays describe the geometry information between the two cameras to ensure that the epipolar constraints are satisfied. As a result, these two types of constraints/errors are measured under the same 3D coordinate system. Therefore, the epipolar constraints and the reconstruction errors are naturally balanced. We have conducted computer simulations and real experiments, and the results show that the proposed calibration method is effective in improving the accuracy compared with the state-of-the-art methods.
- 28 November 2024
- Published by Elsevier in Precision Engineering
- Vol. 92, 30-38
Phase-shifting profilometry for 3D shape measurement of moving objects on production lines
- 28 November 2024
- Published by Elsevier in Precision Engineering
- Vol. 92, 30-38
Phase-shifting profilometry for 3D shape measurement of moving objects on production lines
- 29 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 73 (00189456) , 1-11
Three-Dimensional Surface-Shape Measurement of Moving Billets Using Phase-Shifting Profilometry
Three-dimensional shape of the billet/slab surface is an import cue to detect the surface defects, but it is difficult for phase-shifting profilometry to handle moving objects with textured surface such as casting billets/slabs. There are two issues that need to be addressed together, including object tracking and phase error correction. We propose a method to deal with these two issues. A projection and imaging technique is introduced to track the objects covered with fringe patterns. Based on object tracking, we find that phase errors are mainly induced by rotations and translations of the measured object and we propose a model to describe the phase error distributions. The coefficients of the model are determined by actual phase error distributions, and algorithms for obtaining actual phase error distributions are developed. Phase error distributions provided by the proposed model with determined coefficients are used for phase error correction. Simulations and experiments are carried out to validate the proposed method. Via the proposed method, the three-dimensional reconstruction result of the billet surface measured in moving conditions can provide stereoscopic details as that measured in a stationary state.
- 29 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 73 (00189456) , 1-11
Three-Dimensional Surface-Shape Measurement of Moving Billets Using Phase-Shifting Profilometry
Three-dimensional shape of the billet/slab surface is an import cue to detect the surface defects, but it is difficult for phase-shifting profilometry to handle moving objects with textured surface such as casting billets/slabs. There are two issues that need to be addressed together, including object tracking and phase error correction. We propose a method to deal with these two issues. A projection and imaging technique is introduced to track the objects covered with fringe patterns. Based on object tracking, we find that phase errors are mainly induced by rotations and translations of the measured object and we propose a model to describe the phase error distributions. The coefficients of the model are determined by actual phase error distributions, and algorithms for obtaining actual phase error distributions are developed. Phase error distributions provided by the proposed model with determined coefficients are used for phase error correction. Simulations and experiments are carried out to validate the proposed method. Via the proposed method, the three-dimensional reconstruction result of the billet surface measured in moving conditions can provide stereoscopic details as that measured in a stationary state.
- 20 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 73 (00189456) , 1-13
An AttnGAN-Based Macro Image Synthesis Model for Predicting the Internal Structure of Billets
The internal structure of casting billets directly affects the performance of steel products, but in practice, it can not be detected online. Macro images (low-magnification snapshots of the casting billet sample cross-sections) can visualize the structure, but obtaining them is lagging and costly. Numerical simulation methods are time-consuming to calculate, and only corrected for a few steel grades, making accuracy difficult to guarantee under other working conditions. Thus, we propose a model based on generative adversarial networks (GANs) to quickly predict the internal structure of billets, which can synthesize macro images from production parameters. However, the macro images are less visually varied and have small and dense features. Conventional GANs struggle to generate macro images of casting billets. Therefore, this model is based on the attentional GAN (AttnGAN) to learn different level features through the supervision of multi-resolution images. The generators apply three-branch residual blocks to predict new pixels by multi-scale information, which improves the realism of details. Additionally, considering the ordering property of continuous casting, a series model of the deep neural network and recurrent neural network is designed to encode input parameters. It is combined with image information pretraining to simplify the learning of the mapping relationship in the macro image synthesis model. After comparison with other text-to-image generation models, our model performs better in terms of evaluation metrics and visual effects. The prediction results are verified using two samples from SCM435-M steel. The obtained results have a high consistency with the real data, with the absolute error of the equiaxed crystal rate being 0.84% and 0.6%, respectively. Moreover, the inference speed of our model is fast, which is of reference value for the optimization of the continuous casting process.
- 20 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 73 (00189456) , 1-13
An AttnGAN-Based Macro Image Synthesis Model for Predicting the Internal Structure of Billets
The internal structure of casting billets directly affects the performance of steel products, but in practice, it can not be detected online. Macro images (low-magnification snapshots of the casting billet sample cross-sections) can visualize the structure, but obtaining them is lagging and costly. Numerical simulation methods are time-consuming to calculate, and only corrected for a few steel grades, making accuracy difficult to guarantee under other working conditions. Thus, we propose a model based on generative adversarial networks (GANs) to quickly predict the internal structure of billets, which can synthesize macro images from production parameters. However, the macro images are less visually varied and have small and dense features. Conventional GANs struggle to generate macro images of casting billets. Therefore, this model is based on the attentional GAN (AttnGAN) to learn different level features through the supervision of multi-resolution images. The generators apply three-branch residual blocks to predict new pixels by multi-scale information, which improves the realism of details. Additionally, considering the ordering property of continuous casting, a series model of the deep neural network and recurrent neural network is designed to encode input parameters. It is combined with image information pretraining to simplify the learning of the mapping relationship in the macro image synthesis model. After comparison with other text-to-image generation models, our model performs better in terms of evaluation metrics and visual effects. The prediction results are verified using two samples from SCM435-M steel. The obtained results have a high consistency with the real data, with the absolute error of the equiaxed crystal rate being 0.84% and 0.6%, respectively. Moreover, the inference speed of our model is fast, which is of reference value for the optimization of the continuous casting process.
- 8 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 73 (00189456) , 1-16
Online Recognition of Manually Spray-Printed Ladle Numbers in Complex Dynamic Environments
During the converter process, it is crucial to automatically identify and record ladle numbers to track steel product quality and enhance automation levels. However, the steelmaking environment presents several challenges, including intricate ladle scheduling, varying lighting conditions, severe background interference, and significant disparities between manually spray-printed ladle number characteristics and publicly available datasets. The combination of these problems makes it challenging to perform accurate and real-time ladle number identification. In response, this article suggests an automatic ladle number recognition approach based on deep learning and image processing. First, a double-region object detection model based on YOLOv5 is employed to capture keyframe images of the ladle to be identified from the video stream. Then, a method that can enable the acquisition of an accurate region of ladle numbers in sophisticated industrial settings is proposed to address the distortion of numerical features caused by lighting variations and background interference in industrial environments. Last, leveraging the proprietary dataset found and a ladle number recognition model integrating CNN and multiframe image fusion is designed, developing multithreading design and image queue management to ensure real-time and accurate ladle number recognition. In this study, the video data of a steel plant is used for testing. Through testing 176 steelmaking production cycles, all ladle numbers are accurately identified prior to finishing charging molten iron, indicating the high accuracy and real-time capability of the recognition system.
- 8 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 73 (00189456) , 1-16
Online Recognition of Manually Spray-Printed Ladle Numbers in Complex Dynamic Environments
During the converter process, it is crucial to automatically identify and record ladle numbers to track steel product quality and enhance automation levels. However, the steelmaking environment presents several challenges, including intricate ladle scheduling, varying lighting conditions, severe background interference, and significant disparities between manually spray-printed ladle number characteristics and publicly available datasets. The combination of these problems makes it challenging to perform accurate and real-time ladle number identification. In response, this article suggests an automatic ladle number recognition approach based on deep learning and image processing. First, a double-region object detection model based on YOLOv5 is employed to capture keyframe images of the ladle to be identified from the video stream. Then, a method that can enable the acquisition of an accurate region of ladle numbers in sophisticated industrial settings is proposed to address the distortion of numerical features caused by lighting variations and background interference in industrial environments. Last, leveraging the proprietary dataset found and a ladle number recognition model integrating CNN and multiframe image fusion is designed, developing multithreading design and image queue management to ensure real-time and accurate ladle number recognition. In this study, the video data of a steel plant is used for testing. Through testing 176 steelmaking production cycles, all ladle numbers are accurately identified prior to finishing charging molten iron, indicating the high accuracy and real-time capability of the recognition system.
- Open Access
Automatic Detection of Cast Billet Dendrite Based on Improved Hough Transform
Primary dendrite information is one of the most important metrics to measure the quality of continuous cast slabs. The contrast of low magnification images is very low under the influence of illumination and sampling devices, so the traditional dendrite detection method has the problem of missed detections. We propose an automatic dendrite detection method based on an improved Hough transform, which effectively improves the accuracy and efficiency of primary dendrite detection. By using the local grayscale features of the image, a genetic algorithm-based local contrast enhancement algorithm is proposed. Compared with the traditional contrast enhancement algorithm, it can retain all the information of the dendrites. Combined with the image binarization method based on Hessian matrix, we can obtain more detailed information about the dendrites. According to the continuity and solidification characteristics of dendrites, the Hough transform is improved to extract dendrite information, which effectively reduces the computational cost of the Hough transform. The experimental results show that the method of this paper has versatility, and the error is four pixels compared with the manual method, which can provide a reliable basis for the subsequent judgement of the quality of cast billets.
Open AccessAutomatic Detection of Cast Billet Dendrite Based on Improved Hough Transform
Primary dendrite information is one of the most important metrics to measure the quality of continuous cast slabs. The contrast of low magnification images is very low under the influence of illumination and sampling devices, so the traditional dendrite detection method has the problem of missed detections. We propose an automatic dendrite detection method based on an improved Hough transform, which effectively improves the accuracy and efficiency of primary dendrite detection. By using the local grayscale features of the image, a genetic algorithm-based local contrast enhancement algorithm is proposed. Compared with the traditional contrast enhancement algorithm, it can retain all the information of the dendrites. Combined with the image binarization method based on Hessian matrix, we can obtain more detailed information about the dendrites. According to the continuity and solidification characteristics of dendrites, the Hough transform is improved to extract dendrite information, which effectively reduces the computational cost of the Hough transform. The experimental results show that the method of this paper has versatility, and the error is four pixels compared with the manual method, which can provide a reliable basis for the subsequent judgement of the quality of cast billets.
- 10 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 72 (00189456) , 1-12
A Unified and Accurate Subpixel Detector of Control Points for Camera Calibration
Camera calibration is a necessary step in 3D computer vision, and the accuracy of calibration is largely dependent on the localization precision of the control points of the calibration patterns. However, the sub-pixel localization methods for the commonly used two types of control points, corners of the squares and centers of the circles, are markedly different from each other. In this work, we propose a unified approach to sub-pixel localization for the two types of control points. The approach is derived from theoretical analysis and consists of a closed-form solution, followed by algorithms for practical implementation. Both synthetic dataset and real calibration experiments have been used to test the proposed approach and good results have been obtained. Compared with state-of-the-art techniques, the proposed approach provides a unified and high-precision sub-pixel localization method of the control points for the commonly used calibration patterns.
- 10 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 72 (00189456) , 1-12
A Unified and Accurate Subpixel Detector of Control Points for Camera Calibration
Camera calibration is a necessary step in 3D computer vision, and the accuracy of calibration is largely dependent on the localization precision of the control points of the calibration patterns. However, the sub-pixel localization methods for the commonly used two types of control points, corners of the squares and centers of the circles, are markedly different from each other. In this work, we propose a unified approach to sub-pixel localization for the two types of control points. The approach is derived from theoretical analysis and consists of a closed-form solution, followed by algorithms for practical implementation. Both synthetic dataset and real calibration experiments have been used to test the proposed approach and good results have been obtained. Compared with state-of-the-art techniques, the proposed approach provides a unified and high-precision sub-pixel localization method of the control points for the commonly used calibration patterns.