Human detection based on the generation of a background image by using a far-infrared light camera

Area of Science:

• Computer Vision
• Image Processing
• Thermal Imaging

Background:

• Visible light cameras face limitations in human detection due to illumination variations and shadows.
• Thermal cameras offer an alternative but struggle with low resolution, noise, and background temperature fluctuations.

Purpose of the Study:

• To propose a novel method for robust human detection in thermal camera images.
• To address the limitations of existing thermal imaging human detection techniques.

Main Methods:

• A background image is generated using median and average filtering, with subsequent filtering to remove human areas.
• Candidate human regions are identified by comparing input and background images, with adaptive thresholding and noise reduction.
• Detected regions are refined using histogram analysis, size, and aspect ratio information, considering camera perspective.

Main Results:

• The proposed method effectively detects human areas in thermal images.
• Experimental results demonstrate superior performance compared to existing methods across diverse datasets.

Conclusions:

• The novel method significantly enhances human detection accuracy using thermal imaging.
• This research provides a more reliable solution for surveillance and monitoring systems.