IDE: Image Dehazing and Exposure Using an Enhanced Atmospheric Scattering Model

Area of Science:

• Computer Vision
• Image Processing
• Remote Sensing

Background:

• The atmospheric scattering model (ASM) is widely used for hazy image processing but suffers from a dim effect in recovered images.
• An intrinsic limitation of ASM leads to reduced brightness and contrast in dehazed images, hindering accurate scene representation.

Purpose of the Study:

• To address the dim effect in hazy image restoration by introducing an enhanced atmospheric scattering model (EASM).
• To develop a novel, efficient, and robust image dehazing technique (IDE) based on the EASM for improved visibility of outdoor scenes.

Main Methods:

• An enhanced atmospheric scattering model (EASM) was developed by incorporating a light absorption coefficient parameter.
• A gray-world-assumption-based image enhancement technique (IDE) was designed utilizing the EASM.
• IDE employs a global stretch strategy to avoid common image restoration artifacts.

Main Results:

• The proposed IDE effectively eliminates the dim effect observed in traditional ASM-based methods.
• Experimental results demonstrate excellent dehazing performance, superior to state-of-the-art techniques in both quality and efficiency.
• IDE achieved significant improvements in visibility without requiring scene depth information or training data.

Conclusions:

• The developed EASM and IDE provide a robust and efficient solution for hazy image enhancement.
• IDE offers a fast and reliable method for improving the visual quality of outdoor hazy scenes.
• The proposed approach overcomes limitations of existing methods, offering practical advantages for real-world applications.