Fuzhou Du

Prof. Fuzhou Du


Research Focus & Metrics

Academic Career Length

12Years (2015-Present)

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

Representative Publications

View All Publications
30 July 2015 in Journal of Intelligent Manufacturing
Verified
Journal Article
Closed access
PDF file
ZCZhehan Chen
Zhehan Chen
FDFuzhou Du
Fuzhou Du
XTXiaoqing Tang
Xiaoqing Tang

8 August 2019 in Measurement Science and Technology
Verified
Journal Article
Closed access
PDF file
DZDelong Zhao
Delong Zhao
FKFeifei Kong
Feifei Kong
FDFuzhou Du
Fuzhou Du
As a non-contact inspection approach, vision technology usually undertakes the task of positioning, measuring and defect identification in the field of industrial automation. However, traditional visual programs at a high price are often designed for only a single category of products. Furthermore, the quantitative measurement tasks in the industry usually require a rigorous visual environment as well as hardware equipment, which implies a lack of generalization. Hence, it is imperative to establish a robust approach to break the barriers of multi-type product inspection, while reducing both system complexity and costs. This paper proposed an adaptive approach that performs inspections of the pins’ position for multi-type connectors. A joint strategy of deep neural network and pattern matching based on prior knowledge registration is constructed to achieve rapid positioning of sub-elements arranged in the target. Then, a hierarchical extraction method is designed to analyze features with various appearances and improve the anti-interference of vision-based system. The 3D version of the registration algorithm is embedded into the framework to determine abnormal positions of spatial data without reference. The proposed algorithm demonstrates a successful inspection of a total of 33 types of connectors, significant measurement robustness and adaptivity to the target pose, imaging status and feature diversity.

18 December 2017 in Assembly Automation
Verified
Journal Article
Closed access
DWDian Wu
Dian Wu
FDFuzhou Du
Fuzhou Du
Purpose: In the assembly process of the satellite, there will be multiple installation and disassembly operations for the solar wing and the main satellite body (or simulator). However, the traditional method of orientation adjustment by theodolite and two-axis turntable is difficult to coordinate three rotation angles of yaw, pitch and roll, which leads to the complexity of actual operation and dependency on manual experience. Therefore, this paper aims to propose a new method to achieve rapid and precise orientation adjustment. Design/methodology/approach: The similarity relation of the orientation variation matrix in a different coordinate system is studied, and a mapping model of the similarity relation is established. By using multiple element matrices to construct the original rotation matrix, the mapping is solved in quaternion form. Taking the theodolite as a measuring instrument and the Stewart platform as a control equipment, an experiment on installing the solar wing is performed to validate the effectiveness of the algorithm. Findings: Based on the solving algorithm, the orientation adjustment process is simplified to a three-step fixed mode, which is three adjustments to get the parameter of the mapping model, one to adjust the component in place and another to further fine tuning. The final orientation deviation is less than 0.003° and close to the level of using a laser tracker, achieving the required accuracy of 0.0115°. Originality/value: This paper reveals the similarity relation of the variation matrix in the process of orientation adjustment and presents a new method to achieve rapid and precise orientation adjustment for the large-scale component.

3 June 2019 in Assembly Automation
Verified
Journal Article
Closed access
FDFuzhou Du
Fuzhou Du
KWKe Wen
Ke Wen
HYHao Yu
Hao Yu
Purpose: Aiming at the problems of geometric precision misalignment and unconsidered physical constraints between large components during the measurement-assisted assembly, a self-adaptive alignment strategy based on the dynamic compliance center (DCC) is proposed in this paper, using force information to guide alignment compliantly. Design/methodology/approach: First, the self-adaptive alignment process of large components is described, and its geometrical and mechanical characteristics are analyzed based on six-dimensional force/torque (F/T). The setting method of DCC is studied and the areas of DCC are given. Second, the self-adaptive alignment platform of large components driven by the measured six-dimensional F/T is constructed. Based on this platform, the key supporting technologies, including principle of self-adaptive alignment, coordinate transfer, calculation of six-dimensional F/T and alignment process control, are illustrated. Findings: Using the presented strategy, the position and orientation of large component is adjusted adaptively responding to measured six-dimensional F/T and the changes of contact states are consistent with the strategy. Through the setting of DCC, alignment process runs smoothly without jamming. Practical implications: This strategy is applied to the alignment experiment of large components muff coupling. The experimental results show that the proposed alignment strategy is correct and effective and meets the real-time requirement. Originality/value: This paper proposed a novel way to apply force information in large component self-adaptive alignment, and the setting method of DCC was presented to make the alignment process more feasible.

19 January 2023 in Advanced Intelligent Systems
Verified
Journal Article
Open access
FDFuzhou Du
Fuzhou Du
FKFeifei Kong
Feifei Kong
DZDelong Zhao
Delong Zhao
Vision‐based pose estimation is a basic task in many industrial fields such as bin‐picking, autonomous assembly, and augmented reality. One of the most commonly used pose estimation methods first detects the 2D pose keypoints in the input image and then calculates the 6D pose using a pose solver. Recently, deep learning is widely used in pose keypoint detection and performs excellent accuracy and adaptability. However, its over‐reliance on sufficient and high‐quality samples and supervision is prominent, particularly in the industrial field, leading to high data cost. Based on domain adaptation and computer‐aided‐design (CAD) models, herein, a virtual‐to‐real knowledge transfer method for pose keypoint detection to reduce the data cost of deep learning is proposed. To address the disorder of knowledge flow, a viewpoint‐driven feature alignment strategy is proposed to simultaneously eliminate interdomain differences and preserve intradomain differences. The shape invariance of rigid objects is then introduced as constraints to address the large assumption space problem in the regressive domain adaptation. The multidimensional experimental results demonstrate the superiority of the method. Without real annotations, the normalized pixel error of keypoint detection is reported as 0.033, and the proportion of pixel errors lower than 0.05 is up to 92.77%.

14 September 2018 in The Visual Computer
Verified
Journal Article
Closed access
PDF file
DZDelong Zhao
Delong Zhao
FDFuzhou Du
Fuzhou Du