A differential correction based shadow removal method for real-time monitoring

Area of Science:

• Computer Vision
• Image Processing

Background:

• Shadow removal is critical for motion object surveillance and automatic control.
• Existing pixel-based methods struggle with diverse motion patterns between objects and shadows.
• Computational efficiency is a key constraint for real-time monitoring systems.

Purpose of the Study:

• To propose a novel and simple shadow removal method.
• To address the limitations of current methods in distinguishing shadows from objects.
• To improve shadow removal performance in real-time surveillance applications.

Main Methods:

• A novel shadow removal method based on differential correction calculation.
• Analysis of pixel value differences across Red, Green, and Blue channels.
• Leveraging the characteristic weakening trends in shadow areas across color channels.

Main Results:

• The proposed method effectively distinguishes shadow areas from object areas.
• Experiment results show significant improvements in shadow removal performance.
• Average object and shadow detection accuracies improved by at least 10% compared to state-of-the-art methods.

Conclusions:

• The differential correction method offers a simple yet effective approach to shadow removal.
• The technique accurately identifies shadows based on light attenuation patterns.
• This method enhances the reliability of motion object surveillance systems.