Point source illumination simulation and adaptive pixel-level correction for tunnel scenarios.

Area of Science:

• Engineering
• Computer Vision
• Image Processing

Background:

• Tunnels are vital infrastructure requiring high-quality images for operation and maintenance (O and M).
• Point source illumination is common for tunnel imaging but causes non-uniformity (bright centers, dark edges).

Purpose of the Study:

• To develop an adaptive pixel-level illumination correction algorithm for tunnel images.
• To mitigate the bright-center and dark-edge problem in tunnel imagery.

Main Methods:

• Integration of Entropy Measure of Enhancement (EME) with a Gaussian filter for multi-scale non-uniform illumination restoration.
• An EME-based block partitioning strategy and adaptive Gaussian filter cutoff frequency determination based on local brightness.
• Development of a point source illumination simulation for dataset creation.

Main Results:

• Effective mitigation of the bright-center and dark-edge illumination problem.
• Demonstrated superiority of the proposed correction algorithm over mainstream methods through qualitative and quantitative experiments.
• Validation using a dedicated tunnel image dataset generated via simulation.

Conclusions:

• The proposed adaptive algorithm significantly improves tunnel image quality by correcting non-uniform illumination.
• The method offers a practical solution for enhancing tunnel imaging in urban transportation and underground engineering.
• The developed simulation and dataset aid in robust algorithm validation for tunnel O and M.