Related Experiment Video
Updated: Jun 12, 2026
08:39
Shaping the Amplitude and Phase of Laser Beams by Using a Phase-only Spatial Light Modulator
Published on: January 28, 2019
Laser-induced glare effect and image reconstruction via a frequency-aware transformer
Optics Express
|June 11, 2026
Summary
Researchers developed LaserFormer, a novel AI network, to restore images damaged by laser glare. This technology effectively reconstructs visual information lost to laser interference, improving optoelectronic imaging systems for applications like autonomous driving and security.
Area of Science:
- Optoelectronics and Imaging
- Artificial Intelligence
- Computer Vision
Background:
- Laser glare poses a significant threat to optoelectronic imaging systems by saturating pixels and obscuring scene information.
- Existing algorithms struggle to reconstruct images affected by complex laser patterns, leading to information loss and color distortion.
- Laser glare exhibits unique characteristics, including high-intensity speckle noise, distinct from natural light glare.
Purpose of the Study:
- To experimentally investigate the laser-induced glare effect on electro-optical imaging under various jamming configurations.
- To propose and evaluate a novel deep learning network for restoring laser-glared images.
- To establish a real-world dataset of laser-glared images for benchmarking.
Main Methods:
- Experimental analysis of laser glare effects, including image degeneration patterns (occlusion, halo) and noise characteristics.
- Development of LaserFormer, a frequency-aware transformer network with a dual-path strategy for spatial and frequency information processing.
- Integration of a sparse feature extractor within LaserFormer to capture high-spatial-frequency details and structural information.
Main Results:
- Laser glare causes hybrid image degeneration with occlusion and halos, resulting in information loss and color shifts.
- LaserFormer demonstrated superior performance over existing CNN and transformer models in restoring laser-glared images.
- Quantitative improvements include 1.43% PSNR, 0.83% SSIM, 13.37% FID reduction, and 14.37% LPIPS reduction.
Conclusions:
- LaserFormer effectively mitigates laser glare distortions, enhancing both visual quality and contextual accuracy of reconstructed images.
- The proposed method offers a deployable solution for improving optoelectronic imaging system robustness in challenging lighting conditions.
- The developed dataset and LaserFormer advance the field of laser glare mitigation for applications in autonomous driving, security, and reconnaissance.
