Related Experiment Video
Updated: Jun 9, 2025
13:19
Deep Neural Networks for Image-Based Dietary Assessment
Published on: March 13, 2021
9.0K
An Adaptive Parameter Optimization Deep Learning Model for Energetic Liquid Vision Recognition Based on Feedback
Lu Chen1, Yuhao Yang1, Tianci Wu1
1School of Aeronautics and Astronautics, University of Electronic Science and Technology of China, Chengdu 611731, China.
Sensors (Basel, Switzerland)
|October 26, 2024
Summary
A new deep learning model, DBN-AGS-FLSS, precisely detects liquid flow and viscosity in real-time. This computer vision approach enhances industrial monitoring accuracy for diverse and reflective liquids.
Area of Science:
- Computer Vision
- Deep Learning
- Industrial Monitoring
Background:
- Precise liquid flow and viscosity detection is vital for industrial and environmental monitoring.
- Traditional methods face challenges with diverse liquid samples and reflective properties of energetic liquids.
Purpose of the Study:
- To develop a novel, high-precision, real-time liquid surface pointer detection model.
- To address complexities from sample diversity and reflective properties using computer vision and deep learning.
Main Methods:
- Proposed the DBN-AGS-FLSS integrated deep learning model.
- Combined Deep Belief Networks (DBN), Feedback Least-Squares SVM (FLSS), and Adaptive Genetic Selectors (AGS).
- Utilized bilateral filtering, adaptive contrast enhancement, and a feedback mechanism for parameter optimization.
Main Results:
- Achieved high performance metrics: 99.37% accuracy, 99.36% precision, 99.16% F1 score, and 99.36% recall.
- Demonstrated a rapid inference speed of 1.5 ms/frame.
- Validated superior performance in complex detection scenarios.
Conclusions:
- The DBN-AGS-FLSS model offers practical and reliable liquid detection.
- Opens new avenues for real-time industrial monitoring and automated systems.
- Provides a valuable reference for future computer vision detection technologies.
