Infrared Target Detection Based on Joint Spatio-Temporal Filtering and L1 Norm Regularization

Area of Science:

• Computer Vision
• Signal Processing
• Infrared Imaging

Background:

• Complex backgrounds in infrared imagery lead to high false alarm rates and low target recognition.
• Existing methods struggle with dynamic background changes and intricate scene elements.

Purpose of the Study:

• To propose a robust principal component decomposition model for effective infrared target detection.
• To suppress complex backgrounds and improve target recognition accuracy.
• To address challenges posed by dynamic backgrounds and spatial complexities.

Main Methods:

• A robust principal component decomposition model incorporating joint spatial and temporal filtering.
• Anisotropic Gaussian kernel diffusion for spatial domain target-background differentiation.
• An inversion model with temporal information and L1 norm regularization for dynamic background suppression.
• L1 norm regularization to characterize the target sparse component.
• Overlapping multiplier method for decomposition and reconstruction.

Main Results:

• The proposed background modeling method demonstrated superior background suppression across diverse scenes.
• Achieved average evaluation index values: SSIM (0.986), BSF (88.357), and IC (18.967).
• The detection method yielded a higher detection rate compared to other algorithms at the same false alarm rate.

Conclusions:

• The developed model effectively suppresses complex backgrounds in infrared target detection.
• The method significantly improves target recognition and reduces false alarms.
• This approach offers a robust solution for infrared target detection in challenging scenarios.