- Advanced Measurement and Detection Methods
- Advanced Vision and Imaging
- Optical measurement and interference techniques
- Advanced Sensor and Control Systems
- Advanced Image and Video Retrieval Techniques
- Image Processing Techniques and Applications
- Robotics and Sensor-Based Localization
- Advanced Algorithms and Applications
- Optical Systems and Laser Technology
- Embedded Systems and FPGA Design
- Remote Sensing and Land Use
- Industrial Vision Systems and Defect Detection
- IoT-based Smart Home Systems
- Image and Signal Denoising Methods
- Air Quality Monitoring and Forecasting
- Video Surveillance and Tracking Methods
- Advancements in Battery Materials
- Emotion and Mood Recognition
- Flow Measurement and Analysis
- Advanced Computational Techniques and Applications
- Image and Video Stabilization
- Advanced Neural Network Applications
- Water Quality Monitoring Technologies
- Advanced Image Fusion Techniques
- Image Enhancement Techniques
Tianjin University of Technology
2015-2024
Zhejiang Environmental Monitoring Center
2024
Harbin Engineering University
2023
Powerchina Huadong Engineering Corporation (China)
2019
Xi'an University of Technology
2016
Tianjin University
2011
Shanghai Institute of Measurement and Testing Technology
2009
Yantai University
2009
Behavior is the first visible change in an animal species after exposure to its own or environmental stressors and a sensitive indicator. Fish are social animals, abnormality of group behavior more indicative about particular event than individual behavior, providing effective informeqation changes. The not only reflected spatial distribution, but also temporal movement changes under influence pressure factors. This paper proposes discrimination method based on convolutional neural network...
Both transformer and one-stage detectors have shown promising object detection results attracted increasing attention. However, the developments in effective domain adaptive techniques still not been widely used. In this paper, we investigate issue propose a novel improved You Only Look Once (YOLO) model based on cross-attention strategy transformer, called CAST-YOLO. This detector is Teacher–Student knowledge transfer-based detector. We design encoder layer (TE-Layer) convolutional block...
In this paper, we investigate an interesting problem, i.e., unsupervised cross-corpus speech emotion recognition (SER), in which the training and testing signals come from two different corpora. Meanwhile, are labeled, while label information of is entirely unknown. Due to setting, (source) (target) may have feature distributions therefore lots existing SER methods would not work. To deal with propose a domain-adaptive subspace learning (DoSL) method for projection matrix can transform...
In many practical applications, a speech emotion recognition model learned on source (training) domain but applied to novel target (testing) degenerates even significantly due the mismatch between two domains. Aiming at learning better for domain, paper investigates this interesting problem, i.e., unsupervised cross-corpus (SER), in which training and testing signals come from different corpora. Meanwhile, are labeled, while label information of is entirely unknown. To deal with we propose...
The quality of the air exerts considerable effects on human health, and meteorological factors affect quality. relationships between parameters are complex dependency correlations. This article is based monitoring data obtained from a station in Binjiang District, Hangzhou City, China, spanning 01:00 14 April 2021 to 23:00 31 December 2021. Informer model was used explore parameters’ response factors. By analyzing 12 different kinds 2-Minute Average Wind Speed (2-MAWSP), 10-Minute...
The kernel of stereo vision system is matching. In this paper proposed, an efficient image matching algorithm based on Speed Up Robust Features (SURF). Compared with traditional methods, approach has enormous advantages less computations and short time-consuming. uses SURF to extract feature points. descriptor the determined by Haar wavelet. RANdom SAmple Consensus (RANSAC) used eliminate false matches wrong match point. experimental results show that robust fast speed. That makes important...
An apparatus for the absolute determination of sound speed in water based on time-of-flight technique is described. The time measurement realized by hardware circuits and distance a double-beam plane-mirror interferometer. A highly accurate chip, with resolution approximately 90 ps, employed measurements. acoustic path length adjustable can be measured directly. Two transducers are used transmitting receiving ultrasonic signals without reflection. immersed thermostatic vessel that maintains...
For some measurement and detection applications based on video (sequence images), if the exposure time of camera is not suitable with motion speed photographed target, fuzzy edges will be produced in image, poor lighting condition aggravate this edge blur phenomena. Especially, existence noise industrial field environment makes extraction become a more difficult problem when analyzing posture high-speed moving target. Because are always both kind high-frequency information, it to make...
With the rapid development of Internet Things, remote intelligent control has become an inevitable trend in field industrial development. To solve problems complicated wiring inverter monitoring system, insufficiently advanced system architecture, and insufficient message push efficiency, this paper proposes a communication protocol based on Message Queuing Telemetry Transport (MQTT) ESP8266 hardware module. The PC terminal starts MQTT cloud server subscribes messages to as client. Once...
Three-dimensional (3D) object detection has a vital effect on the environmental awareness task of autonomous driving scenarios. At present, accuracy 3D significant improvement potential. In addition, point cloud is not uniformly distributed regular grid because its disorder, dispersion, and sparseness. The strategy convolution neural networks (CNNs) for feature extraction limitations potential information loss empty operation. Therefore, we propose graph network (GNN) detector based neighbor...
Because of its advantage panoramic sight with a single compact visual scene, Omni-directional vision technology has extreme application value in many fields. One the most effective ways to establish system is using fisheye lens. However it brings strong unavoidable inherent distortion while provides hemispherical field view. But can be corrected by some image processing techniques. A method for correction such distorted derived this paper. After introduce two kings perspective projection...
A novel calibration method for fisheye lens camera is proposed in this paper. The based on equidistance projection model of lens. inner structure the which consists many pieces so complex that it very difficult and unnecessary to construct its physical model. optical center (u <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</sub> ,v ), virtual focal fish-eye l ration distortion x coordinate Kx Ky y should be calibrate mathematical mode by us....
Conventional stereo vision systems have a small field of view (FOV) which limits their usefulness for certain applications. While panorama is able to "see" in all directions the observation space, scene depth information missed because mapping from 3D reference coordinates 2D panoramic image. In this paper, we present an innovative system builds by special combined fish-eye lenses module, and capable producing coordinate whole global space acquiring no blind area 360°×360° image...