Choosing the filter for catenary image enhancement method based on the non-subsampled contourlet transform

Area of Science:

• Electrical Engineering
• Image Processing
• Computer Vision

Background:

• Catenary fault diagnosis relies heavily on image processing techniques.
• Low-contrast catenary images hinder accurate fault detection.
• Existing methods like Contourlet Transform (CT) can introduce artifacts.

Purpose of the Study:

• To develop an effective image enhancement method for low-contrast catenary images.
• To improve the performance of catenary fault diagnosis systems.
• To address artifact issues in enhanced catenary images.

Main Methods:

• Utilizing the Non-subsampled Contourlet Transform (NSCT) for image decomposition.
• Applying a nonlinear enhancement function to NSCT coefficients.
• Inverting the NSCT to obtain the enhanced image.
• Comparing the proposed method with Lifting Wavelet Transform, Retinex, Dark Channel, Curvelet Transform, and CT.

Main Results:

• The proposed NSCT-based method significantly enhances contrast and clarity of catenary images.
• It effectively eliminates artifacts while preserving crucial image details like edges and textures.
• Quantitative metrics show improved detail variance, signal-to-noise ratio, and edge preservation compared to CT.

Conclusions:

• The combined NSCT and nonlinear enhancement function is an excellent approach for catenary image enhancement.
• This method offers superior performance in improving image quality for fault diagnosis.
• The technique effectively preserves image details and reduces artifacts.