Extended Codebook with Multispectral Sequences for Background Subtraction

Area of Science:

• Computer Vision
• Image Processing
• Remote Sensing

Background:

• Background subtraction is crucial for real-time object detection.
• Traditional Red-Green-Blue (RGB) models have limitations in complex environments.
• The Codebook model is a popular real-time background subtraction technique.

Purpose of the Study:

• To extend the Codebook model for multispectral sequences.
• To introduce a self-adaptive mechanism for improved performance and parameter efficiency.
• To evaluate the effectiveness of the multispectral Codebook model in moving object detection.

Main Methods:

• Extension of the Codebook model from RGB to multispectral data.
• Development of a self-adaptive mechanism using statistical information.
• Introduction of Spectral Information Divergence, Brightness, and Spectral Distortion for evaluation.

Main Results:

• The multispectral self-adaptive Codebook model demonstrated superior moving object detection capabilities compared to RGB sequences.
• The self-adaptive mechanism improved performance and reduced the need for manual parameter tuning.
• Experiments on diverse multispectral datasets validated the model's robustness.

Conclusions:

• The multispectral self-adaptive Codebook model offers enhanced performance for moving object detection.
• This research framework facilitates future applications of multispectral sequences in this domain.
• The proposed method provides a more robust and efficient solution for real-time background subtraction.