Algorithm Design for Edge Detection of High-Speed Moving Target Image under Noisy Environment

Area of Science:

• Image processing
• Computer vision
• Signal processing

Background:

• Fuzzy edges in images, caused by motion blur and poor lighting, complicate edge detection.
• Noise in industrial environments further degrades image quality, hindering analysis of high-speed moving targets.
• Traditional frequency-based methods struggle to differentiate between high-frequency noise and edge information.

Purpose of the Study:

• To develop a noise-tolerant edge detection method for images with fuzzy edges.
• To directly extract edge information from noisy images without full image restoration.
• To improve the analysis of high-speed moving targets in industrial applications.

Main Methods:

• A novel edge detection method utilizing the correlation of wavelet transform coefficients across layers.
• Incorporation of neural network activation function principles for wavelet coefficient judgment.
• Design of rational coefficients wavelet filters (length 8-4 biorthogonal) with reduced vanishing moments.

Main Results:

• The proposed method successfully extracts edge information from noisy images with fuzzy edges.
• Demonstrated superiority over existing methods in comparative experiments.
• Effective handling of trade-offs between noise and edge signals using wavelet coefficient characteristics.

Conclusions:

• The developed wavelet-based method offers a robust solution for edge detection in challenging industrial imaging scenarios.
• The technique provides clear edge detection for high-speed moving targets with motion-induced blur.
• This approach enhances the reliability of measurement and detection applications relying on image analysis.