Polarized image near-natural color fusion algorithm for target detection

Area of Science:

• Computer Vision
• Image Processing
• Remote Sensing

Background:

• Target detection is crucial for automatic systems, often relying on visible imagery.
• Polarization information has been shown to effectively highlight man-made targets.

Purpose of the Study:

• To develop and evaluate an algorithm that fuses polarized and visible images to enhance target detection accuracy.
• To improve the gradient and contrast of fused images for better visual analysis.

Main Methods:

• Simultaneous conversion of polarization and visible images to HSV color space.
• Transformation to Lab color space for color transfer after initial fusion.
• Segmentation using bisecting k-means algorithm for region-based color transfer.
• Final fusion image transformation back to RGB color space.

Main Results:

• Significant improvements in gradient (115%) and contrast (235.3%) compared to visible images.
• The fused image is suitable for naked-eye viewing.
• The proposed algorithm demonstrates a notable enhancement in detection accuracy.

Conclusions:

• The fusion of polarized and visible images offers a substantial improvement for target detection.
• The developed algorithm provides a practical and effective solution for enhancing image analysis in automatic systems.