A5001374887- Water Quality Monitoring Technologies
- Fish Ecology and Management Studies
- High voltage insulation and dielectric phenomena
- Underwater Vehicles and Communication Systems
- Power Transformer Diagnostics and Insulation
- Advanced Neural Network Applications
- Fish biology, ecology, and behavior
- Hydrological Forecasting Using AI
- Image Processing Techniques and Applications
- Plant Water Relations and Carbon Dynamics
- Advanced Image Processing Techniques
- Industrial Vision Systems and Defect Detection
- Aquaculture disease management and microbiota
- Fire Detection and Safety Systems
- Non-Destructive Testing Techniques
- Water Quality Monitoring and Analysis
- Aquaculture Nutrition and Growth
- Identification and Quantification in Food
- Advanced Vision and Imaging
- Image Enhancement Techniques
National Engineering Research Center for Information Technology in Agriculture
2020-2024
Jilin Electric Power Research Institute (China)
2023
Northwest A&F University
2023
East China Jiaotong University
2020
Nanjing Agricultural University
2013
The automatic detection and identification of fish from underwater videos is great significance for fishery resource assessment ecological environment monitoring. However, due to the poor quality images unconstrained movement, traditional hand-designed feature extraction methods or convolutional neural network (CNN)-based object algorithms cannot meet requirements in real scenes. Therefore, realize recognition localization a complex environment, this paper proposes novel composite framework...
In intensive aquaculture, the number of fish in a shoal can provide valuable input for development intelligent production management systems. However, traditional artificial sampling method is not only time consuming and laborious, but also may put pressure on fish. To solve above problems, this paper proposes an automatic counting based hybrid neural network model to realize real-time, accurate, objective, lossless population far offshore salmon mariculture. A multi-column convolution...
In aquaculture, accurate and automatic quantification of fish behaviour can provide useful data input for production management decision-making. recent years, with the focus on welfare, it has become urgent to study use nondestructive quantitative methods in aquaculture. this paper, based literature past 30 nonintrusive are analysed. Firstly, several important behaviours aquaculture listed, is summarized four stages: detection, tracking, feature extraction recognition. Then, quantification,...
The real-time classification of fish feeding behavior plays a crucial role in aquaculture, which is closely related to cost and environmental preservation. In this paper, Fish Feeding Intensity model based on the improved Vision Transformer (CFFI-Vit) proposed, capable quantifying behaviors rainbow trout (Oncorhynchus mykiss) into three intensities: strong, moderate, weak. process outlined as follows: firstly, we obtained 2685 raw images from recorded videos classified them categories:...
In recent years, residual learning has shown excellent performance on convolutional neural network (CNN)-based single-image super-resolution (SISR) tasks. However, CNN-based SISR approaches have focused mainly the design of deep architectures, and rectified linear units (ReLUs) used in these networks hinder shallow-to-deep information transfer. As a result, methods are unable to utilize some shallow information, improving model is difficult. To solve above issues, this paper proposes an...
Abstract In view of the shortcomings traditional dissolved gas analysis technology low diagnostic veracity and intelligence, this paper proposes to use QPSO optimize nuclear argument in support vector machine (SVM), on basis, (DGA) is used diagnosis transformer faults. Firstly, data preprocessed by DGA technology, processed as input amount fault characteristics. Secondly, for optimization core parameters SVM, algorithm combined with training acquisition. Finally, five kinds feature inputs...
Only a few key fish individuals can play dominant role in actual group, therefore, it is reasonable to infer group activities from the relationship between individual actions. However, complex underwater environment, rapid and similar movements are likely cause indistinct action characteristics, as well adhesion of data distribution, difficult actions directly by using graph convolutional network (GCN). Therefore, this article proposes convolution vector calibration (GCVC) for activity...
Aiming at the problem of untimely fire prevention and control due to missed false alarms in traditional early warning system substations, an intelligent classification algorithm based on multi-sensor information fusion is proposed this paper. Different from with a single sensor, firstly, paper combines temperature, CO concentration smoke sensors build multi-sensing layer detection model, which improves sensitivity certain extent. Then, uses support vector machine (SVM) classify warn fires...
With the increasing demand for power supply, it is particularly important to control indoor temperature of substations in summer. In this paper, transformer room studied: firstly, transfer function ambient given, which influence heat loss on output variation fully considered. Secondly, model prediction method used input and changes mathematical model. order achieve objectives, constraint conditions variables are given process. Finally, experiment shows that predictive can a small range,...