The Algorithm of Watershed Color Image Segmentation Based on Morphological Gradient.

Area of Science:

• Computer Vision
• Image Processing
• Pattern Recognition

Background:

• Traditional watershed algorithms suffer from over-segmentation and sensitivity to reflected light.
• Existing methods often struggle with accurate object contour detection in complex lighting conditions.

Purpose of the Study:

• To propose an improved watershed algorithm for color image segmentation.
• To enhance robustness against reflected light and reduce over-segmentation.
• To improve the accuracy and efficiency of object segmentation.

Main Methods:

• Utilizes a morphological gradient in a novel color space to mitigate reflected light interference.
• Reconstructs the gradient image using morphological opening and closing operations.
• Employs the maximum inter-class variance algorithm for automatic thresholding and calibrates the gradient image for watershed segmentation.

Main Results:

• Achieves accurate and continuous target contours, minimizing segmentation regions to align with human perception.
• Effectively suppresses noise and meaningless areas caused by reflected light.
• Demonstrates a 10% increase in operational efficiency compared to region-growing and automatic threshold methods.
• Reports accuracy and recall rates exceeding 0.98.

Conclusions:

• The proposed improved watershed algorithm offers superior performance in color image segmentation.
• It enhances robustness, maintains edge information, and improves applicability.
• The method provides a significant advancement over existing techniques, particularly in challenging image conditions.