A5103183814- Advanced Neural Network Applications
- Autonomous Vehicle Technology and Safety
- Remote Sensing and LiDAR Applications
- Video Surveillance and Tracking Methods
- Advanced Measurement and Metrology Techniques
- Image and Object Detection Techniques
- Image Enhancement Techniques
- Software Reliability and Analysis Research
- Mechanics and Biomechanics Studies
- Design Education and Practice
- Collaboration in agile enterprises
- Robotics and Sensor-Based Localization
- Spectroscopy and Chemometric Analyses
- Robotic Path Planning Algorithms
- Semantic Web and Ontologies
- Hydraulic and Pneumatic Systems
- Advanced Image Fusion Techniques
- Artificial Immune Systems Applications
- Advanced Image Processing Techniques
- Metaheuristic Optimization Algorithms Research
- Vehicle License Plate Recognition
- Machine Fault Diagnosis Techniques
- Scientific Measurement and Uncertainty Evaluation
- Blind Source Separation Techniques
- Optical measurement and interference techniques
Zhengzhou University of Light Industry
2009-2025
Shanghai Tunnel Engineering (China)
2024-2025
Beijing Institute of Technology
2009
Abstract Aiming at addressing the complex lighting conditions in shield construction environments and enhancing contrast of low-light images while preserving image details, this paper presents an enhancement technique based on multi-scale fusion. The proposed method first duplicates input into two copies. Subsequently, copy undergoes color adjustment via adaptive weighted Retinex algorithm for restoration (A-MSRCR) algorithm. Meanwhile, second detail preservation by combining Gamma...
To meet the requirements for automation and precision in segment assembling enhance quality efficiency of assembly, a multi-objective trajectory planning method considering error uncertainty is proposed strategy erector under conditions heavy load low-speed assembly. Based on forward kinematics model erector, working space determined by Monte Carlo method, key track points are according to actual operation requirements. The kinematic mapping relationship between Cartesian joint established...
In urban road scenes, due to the small size of traffic signs and large amount surrounding interference information, current methods are difficult achieve good detection results in field unmanned driving. To address aforementioned challenges, this paper proposes an improved E-YOLOv4-tiny based on YOLOv4-tiny. Firstly, article constructs efficient layer aggregation lightweight block with deep separable convolutions enhance feature extraction ability backbone. Secondly, presents a fusion...
A multi-strategy improved ant colony algorithm is proposed. In order to solve the problem of solving TSP in algorithm, it has problems low solution accuracy, easy fall into local optimum, and efficiency. The nearest neighbor method used influence distribution initial pheromone reduce concentration on short path stage algorithm. Based mutation adjustment transfer rule, a mean cross-evolution strategy combined with enhance global space Ability ability avoid falling optimum. Then, iterative...
The foggy images captured by drones are nonuniform due to inhomogeneous distribution of fog in higher altitude, leading the obvious thickness differences images. This paper proposes a classification guided thick removal network for drone imaging, termed ClassifyCycle. input into proposed module (ICLFn) enhance reliability follow-up learning network. style migration (ISMn) is introduced reduce image distortion, such as hue artifact and texture distort. ClassifyCycle does not require paired...
Abstract Measuring a spherical surface with small segment angle (e.g. 1°) precisely is big challenge in scientific research and engineering applications. To fit sphere surfaces angles high precision, five-parameter constrained nonlinear least square fitting (CNLSF) method presented this work, where an Adam optimization algorithm prominent anti-interference ability employed. We discuss various factors which influence the final uncertainty evaluation of sphere’s radius. A Monte Carlo...
Abstract Product requirement is the goal of product development and design activities, which originates from all stages life cycle. In process acquiring requirements, there are some problems such as incomplete mining requirements fuzzy structural expression. view above problems, this paper proposes a cycle acquisition method based on axiomatic design. Firstly, web technology applied to collect information, HLDA hierarchy model, information classified according whole products; then, model...
As a systematic design method, modular requires modeling techniques to prepare information structure for construction and evolution of product system, which generally include decomposition that specific family optimization. In this paper, multi-view methodology is proposed such an based on relationship constraint network. Firstly, locations are given according the whole close-loop process design, framework method given. Secondly, correlative among elements organized in form network, into...