Image Recognition and Parameter Analysis of Concrete Vibration State Based on Support Vector Machine.

Area of Science:

• Materials Science
• Civil Engineering
• Computer Vision

Background:

• Monitoring concrete integrity is crucial for structural safety.
• Assessing concrete under vibration is challenging due to complex feature extraction.

Purpose of the Study:

• To evaluate the feasibility of using directional gradient histogram and SVM for concrete vibration state identification.
• To analyze the impact of directional gradient histogram parameters on recognition accuracy.

Main Methods:

• Feature extraction from concrete images using directional gradient histogram technology.
• Classification of vibration states using a support vector machine (SVM) model.
• Analysis of parameter sensitivity (block size, statistical intervals, binarization threshold) on recognition accuracy.

Main Results:

• Directional gradient histogram-SVM technology is feasible for concrete vibration state identification.
• Recognition accuracy is sensitive to directional gradient histogram parameters, with an optimal range.
• 100% recognition accuracy was achieved with optimized parameters and high-resolution images (1024x1024 pixels).

Conclusions:

• The proposed method offers a viable approach for non-destructive concrete condition monitoring.
• Parameter optimization is key to maximizing the performance of the directional gradient histogram-SVM technique.
• This technique holds potential for real-world applications in structural health monitoring.