Target detection enhancement method through multi-dimensional information differential projection and fusion

Area of Science:

• Optics and Photonics
• Computer Vision
• Remote Sensing

Background:

• Spectral polarization imaging (SPI) offers rich spatial, spectral, and polarization data for detection.
• Current SPI methods face challenges with large datasets and detecting targets in complex, low-contrast environments.

Purpose of the Study:

• To develop an innovative target detection enhancement method for SPI systems.
• To improve target identification in complex natural scenes with reduced contrast and clutter.

Main Methods:

• Developed a target retrieval algorithm using angle of polarization purification.
• Implemented a novel multi-dimensional information differential projection and fusion strategy.
• Utilized self-developed SPI systems for data acquisition.

Main Results:

• The proposed method effectively exploits multi-dimensional distinctions between targets and backgrounds.
• Successfully detected targets with shielding, camouflage, and shape loss in grasslands, woodlands, and sky.
• The fused image clearly visualizes target-background distinctions, enhancing detection.

Conclusions:

• The method significantly enhances scenario comprehension and detection efficiency.
• It offers advantages for military reconnaissance applications due to low computational requirements and no need for extensive model training.
• This approach provides a robust solution for target detection in challenging natural environments.