Jian Liang

Jian Liang

Auto-generated profile, based on publicly available information

Jian Liang

State Key Laboratory of Precision Measurement Technology and Instruments, Tianjin University, Tianjin, China

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
5 September 2024 in IEEE Transactions on Circuits and Systems for Video Technology
Verified
Journal Article
Closed access
ZZZongyang Zhao
Zongyang Zhao
JKJiehu Kang
Jiehu Kang
LFLuyuan Feng
Luyuan Feng
JLJian Liang
Jian Liang
YRYuqi Ren
Yuqi Ren
WBWu Bin
Wu Bin
Partial point cloud registration is an essential and fundamental component of generating complete 3D shapes, aiming at converting partial scans into a unified coordinate system. However, existing point cloud registration methods suffer from inadequately rich local feature extraction and feature interaction. In addition, these methods still face challenges in modeling the global contextual information of point clouds sufficiently, which limits the improvement in registration effectiveness, especially for partial-to-partial point cloud registration with high noise. To overcome these issues, this paper proposes an enhanced PointNet and assignable weights transformer network (LFA-Net) for partial point cloud registration. The model achieves coarse-to-fine point cloud registration through three core modules. First, the Sufficient Local Feature Extraction Module (LFM) is constructed to extract various local feature information. Then, the Adequate Feature Aggregation Module (FAM) is designed to integrate the feature information from different point clouds. Finally, the Assignable Weights Transformer Module (ATM) is presented to stimulate the model’s global modeling ability during the registration process, enabling the selection of representative points for optimal point cloud registration. Extensive experiments conducted on ModelNet40 using partially overlapping point clouds illustrate the superior registration performance of LFA-Net compared with other state-of-the-art methods. Moreover, Numerous experiments on synthetic and real-word datasets further indicate that LFA-Net also has significant advantages in registering partial point clouds with noise and unseen categories, which demonstrates its excellent robustness and generalization ability for real-world practical application.

1 March 2024 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
ZZZongyang Zhao
Zongyang Zhao
JKJiehu Kang
Jiehu Kang
WBWu Bin
Wu Bin
TYTao Ye
Tao Ye
JLJian Liang
Jian Liang
The incursion of railway obstacles poses a serious risk to train operations, and numerous accidents occur during train shunting. However, existing algorithms still struggle with finding a compromise between detection accuracy and speed during train movement. Moreover, their accuracy and robustness are inadequate, specifically when handling small objects in complicated railway scenarios. To overcome these issues, this paper proposes an efficient network using convolution and transformer (AE-Net) for performing accurate and real-time detection of railway obstacles to ensure driving safety. First, the Enhanced and Lightweight Transformer Module (ETM) is constructed to strengthen the model’s global modeling ability. Then, the Lightweight Feature Integration Module (LIM) is presented to integrate multi-branch feature information and reduce model complexity. Finally, the Reinforced Multi-Scale Feature Fusion Module (RFM) is utilized to enhance the multi-scale object detection capability, especially for small obstacles. The presented algorithm realizes 95.29% mAP and 145 FPS on the railway dataset, which is superior to YOLOv5s. In addition, the experiment on MS COCO further shows that AE-Net can perform a considerably better detection than current state-of-the-art models. Hence, it is practicable to employ AE-Net in actual railway and further more complex multi-target scenarios.

30 September 2025 in Measurement
Verified
Journal Article
Closed access
JLJian Liang
Jian Liang
ZSZefeng Sun
Zefeng Sun
JKJiehu Kang
Jiehu Kang
SWShuyang Wang
Shuyang Wang
ZZZongyang Zhao
Zongyang Zhao
BWBin Wu
Bin Wu

24 December 2025 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
SWShuyang Wang
Shuyang Wang
JLJian Liang
Jian Liang
JKJiehu Kang
Jiehu Kang
ZSZefeng Sun
Zefeng Sun
ZZZongyang Zhao
Zongyang Zhao
WBWu Bin
Wu Bin
Accurate calibration of scanning probes is essential for ensuring the measurement precision of coordinate measuring machines (CMMs). However, most existing calibration methods rely on linear models, which are inadequate for compensating nonlinear errors inherent in scanning probes. Furthermore, these methods often overlook the form error of the ruby ball, which further limits calibration accuracy. To address these limitations, this paper proposes a high-precision calibration approach that compensates for both linear and nonlinear errors. For linear errors, we incorporate the wear-induced error of the ruby ball into the parameter identification process, resulting in a more comprehensive and accurate linear calibration model. For nonlinear errors, a novel Nonlinear Error Calibration Neural Network (NEC-NN) is proposed. NEC-NN employs the Regional Modulation Module (RMM) to encode probing direction and the Cascaded Multi-Head Attention Module (C-MHAM) to extract deep-level features, enabling the network to capture directiondependent nonlinear errors. Additionally, normal consistency and spherical surface constraints are integrated into the loss function, reducing the training complexity while improving the adaptability of the model. The CMM used in the experiments has a nominal maximum permissible error of 2 + L[mm]/400 μm. After calibrating the scanning probe using the proposed method, the probe’s maximum probing error on a standard sphere decreased to 0.42 μm, while the CMM’s actual maximum permissible error was reduced to 0.6 + L[mm]/400 μm, demonstrating that the proposed method can significantly improve the overall measurement accuracy of the CMM. The proposed approach provides a more reliable and practical solution for the calibration of CMM scanning probes, thereby enhancing the overall reliability and performance of CMM in industrial applications.

23 June 2025 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
JLJian Liang
Jian Liang
ZSZefeng Sun
Zefeng Sun
JKJiehu Kang
Jiehu Kang
TFTian Feng
Tian Feng
SWShuyang Wang
Shuyang Wang
WBWu Bin
Wu Bin
Geometric errors in coordinate measuring machines (CMMs) significantly degrade measurement accuracy. While laser tracking interferometer (LTI)-based compensation methods are widely used due to their high efficiency, the limited precision of the existing positioning techniques restricts their application in ultrahigh-precision CMMs. To address these challenges, this study introduces an accurate and efficient geometric error compensation method that utilizes LTI positioning for ultrahigh-precision CMMs. The method begins with the development of a geometric error model that employs homogeneous transformation matrices (HTMs) to map end-position deviations to error parameters. A highly accurate and robust positioning algorithm for LTI is then designed, incorporating a two-step process: initial positioning through semidefinite programming (SDP) and fine-tuning using enhanced particle swarm optimization (EPSO). After parameter identification, geometric error compensation is applied based on the established model. The test experiments were conducted on a CMM with the nominal accuracy of $2\:\pm \: L$ [mm]/ $400 \: \:\mu $ m. The positioning results show that the proposed method achieves a distance mean absolute error of 0.10871, demonstrating superior precision over conventional methods. Additionally, the method exhibits excellent robustness and stability under noise interference. After compensation, precision validation results showed that the maximum detection error was reduced to $ 0.26\:\mu $ m, with length measurement errors within $0.5\:\pm \: L$ [mm]/ $400 \:\: \mu $ m. These results highlight substantial improvements in both measurement precision and operational performance. This research presents an effective solution for geometric error compensation in CMMs, offering enhanced performance for industrial measurement applications.

23 May 2025 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
JLJian Liang
Jian Liang
JKJiehu Kang
Jiehu Kang
XWXuesen Wang
Xuesen Wang
ZZZongyang Zhao
Zongyang Zhao
CZChi Zhao
Chi Zhao
WBWu Bin
Wu Bin
Overloaded truck transportation poses a significant threat to road safety. However, mainstream methods for overloaded truck identification suffer from low efficiency, limited monitoring range, and high costs, posing obstacles to curbing overloaded transportation. To tackle these challenges, this study proposes a novel approach called Multi-Aspect Overloaded Truck High-Precision Identification Network (MAOTIN), which achieves accurate identification by analyzing truck trajectory data comprehensively. MAOTIN comprises three efficient and specific modules: overloaded feature identification module (OFIM), trajectory similarity identification module (TSIM), and decision fusion module (DFM). Initially, the trajectory representation method was designed to transform the multidimensional spatiotemporal truck trajectory data into images, aggregating overload features while reducing dimensionality. Subsequently, OFIM and TSIM are developed to comprehensively identify overloaded behavior from typical overloaded features and overloaded trajectories similarity features. Finally, DFM adaptively and fully fuses feature images of different properties to obtain the ultimate identification result. To evaluate the performance of MAOTIN, this study conducted experiments on the constructed overloaded truck dataset. The results indicated that the model achieves an accuracy of 96.4% and 316 FPS, surpassing mainstream image classification networks. Moreover, MAOTIN performed a precision of 97.3% on practical application testing and assisted traffic police in apprehending 7591 overloaded trucks in one quarter, demonstrating its high practical value. Hence, MAOTIN enables accurate, rapid, and comprehensive identification of overloaded trucks and can be effectively applied to the overloaded identification tasks in urban traffic management.