A5025208620- Industrial Vision Systems and Defect Detection
- Optical measurement and interference techniques
- Remote Sensing and Land Use
- Advanced Measurement and Detection Methods
- Image Processing Techniques and Applications
- Optical Systems and Laser Technology
- Remote Sensing and LiDAR Applications
- Digital Imaging for Blood Diseases
- Service-Oriented Architecture and Web Services
- 3D Shape Modeling and Analysis
- Advanced Vision and Imaging
- Autonomous Vehicle Technology and Safety
- Vehicle License Plate Recognition
- Advanced Neural Network Applications
- Traffic Prediction and Management Techniques
- Phytochemistry and Biological Activities
- Smart Agriculture and AI
- Automated Road and Building Extraction
- Video Surveillance and Tracking Methods
- Software Engineering and Design Patterns
- Image and Video Stabilization
- Natural product bioactivities and synthesis
- Sesquiterpenes and Asteraceae Studies
- Web Applications and Data Management
- Retinal Imaging and Analysis
Zhejiang Sci-Tech University
2010-2025
Zhejiang University
2010-2012
Zhejiang University of Science and Technology
2012
Taiyuan University of Technology
2011
Accurate and timely lane detection is imperative for the seamless operation of autonomous driving systems. In this study, leveraging gradual variation features within a defined range width length, we introduce an enhanced Spatial-Temporal Recurrent Neural Network (SCNN) framework. This framework serves as cornerstone innovative hybrid spatial-temporal model detection, which tailored to address prevalent issues substandard performance insufficient real-time processing in intricate scenarios,...
Aiming at the current problems of low detection accuracy high-voltage cabinet switches and large models that are difficult to deploy, a switch method based on improved EfficientNet-YOLOv5s model is proposed. Firstly, offline data expansion online Mosaic enhancement performed switchgear images obtained by video frame extraction alleviate unbalanced sample categories small number samples. Secondly, backbone network YOLOv5s replaced with lightweight EfficientNet-B0 network, original convolution...
Feature extraction and matching is the key process of motion estimation, determines performance video stabilization to a great extent. A novel approach was proposed based on multi-scale colored local invariant features. The transformed image from RGB color model model, built up space Gaussian pyramids, then extracted FAST feature points in multiscale matched by building Fast Retina Key-point (FREAK) descriptors, finally estimated interframe motions M-estimator Sample Consensus (MSAC)...
A new caryolane sesquiterpene, 1β-methoxycaryol-9-one (1), along with eight known compounds (2-9), was isolated from the whole plant of Heteropappus altaicus. The structures were elucidated on basis comprehensive 1D- and 2D-NMR analysis high-resolution mass spectrometry. 13C-NMR data compound 2 are reported for first time.
A new objective fabric pilling detection method is proposed for monochrome without pattern. Firstly, through wavelet multi-resolution decomposition and reconstruction, remove the high-frequency noise such as fluff which significantly different from size of pills low-frequency uneven illumination. After image, analyzing energy change law between each layer, then judged whether it rough or fine texture. Then, according to color characteristics texture fabrics, even symmetry two-dimensional...
The paper introduces the basic steps to create UML use case model with Banking Loans System as an example. On basis of model, capture and analysis system classes from description by grammar analysis. Then through establishing associations for any two creates class diagram System. It provides a reference achieve.
Mean Shift algorithm, a statistic iterative procedure, is robust when applied to farmland image segmentation. It can effectively overcome the influence of shadow, weeds or illumination changes, etc. However, procedure has relatively high time complexity and not meet requirements real-time processing. Based on pyramid we obtain low resolution representation images being processed. Then, run shift algorithm set seed points that selected in image. Through this method, consumption significantly...
Support Vector Machine, a statistic procedure, is robust and has good performance when applied to farmland image segmentation. It can effectively identify the crop rows, even though they have some intersection at points because of weeds leaves growth crops. However, Machine relatively high time complexity cannot meet requirements real-time processing. Based on Pyramid algorithm, we obtain low-resolution representation images being processed. Then training testing are in only. Through this...
Fabric pilling image acquisition process will inevitably cause noise. In order to effectively remove the noise from of pilled fabric obtained by circular locus method, this paper tried use Otsu segmentation, edge detection algorithm and Hough transform extract area, aimed at problems existing in current removal methods, such as texture residual, shadow sensitive, or instability effect removal, Gabor filter textures. The can make full not only periodicity but also directionality texture....
Most nonrigid motions use shape-based methods to solve the problem; however, of discrete cosine transform trajectory-based motion problem is also very prominent. The signal undergoes transformation due characteristics transform. correlation data well extracted such that a better compression achieved. However, it important select number and sequence trajectory basis appropriately. error reconstruction operational costs will increase for high value K (number basis). On other hand, lower would...