Optimized image fusion rule based on targeted image evaluation metrics for vehicle imaging systems

Area of Science:

• Computer Vision
• Image Processing
• Sensor Fusion

Background:

• Perception systems for vehicle assistance degrade in adverse conditions like haze.
• Effective image processing is crucial for reliable autonomous driving and surveillance.

Purpose of the Study:

• To develop an integrated visible polarization and infrared dual-band imaging system.
• To propose an innovative image fusion rule for enhanced perception in complex environments.

Main Methods:

• Integrated visible polarization and infrared dual-band imaging system.
• Novel image fusion rule based on target image evaluation metrics.
• Experimental validation of the fusion method.

Main Results:

• Significantly enhanced texture details in fused images.
• Improved target discrimination capabilities.
• Demonstrated versatility across fusion frameworks and multimodal applications.

Conclusions:

• The proposed system and fusion rule effectively address perception degradation in haze.
• The method offers valuable solutions for assisted driving, remote sensing, and military reconnaissance.
• The fusion rule's versatility supports broader applications in multimodal imaging.