Sand dust image visibility enhancement algorithm via fusion strategy.

Area of Science:

• Computer Vision
• Image Processing

Background:

• Outdoor images in sand dust weather exhibit poor contrast and color distortion.
• These image quality issues hinder the performance of intelligent information processing systems.

Purpose of the Study:

• To propose a novel enhancement algorithm for sand dust-affected images.
• To improve the performance of intelligent systems by enhancing image quality.

Main Methods:

• A fusion strategy combining two sequential components.
• Sand removal using an improved Gaussian model-based color correction algorithm.
• Dust elimination utilizing a residual-based convolutional neural network (CNN).

Main Results:

• The proposed fusion strategy effectively corrects the yellowing hue caused by sand.
• The algorithm successfully removes dust haze disturbance from images.
• Experimental results demonstrate superior performance compared to existing methods.

Conclusions:

• The developed fusion strategy offers a constructive approach for sand dust image enhancement.
• This method provides a foundation for future advancements in the field.
• Enhanced image quality supports more reliable intelligent information processing.