Researchers from Washington State University have developed a new method to increase the clarity of images in conditions of limited air transparency, such as fog or smoke. The method uses reality-reproducing neural networks to analyze images captured in low-visibility conditions and automatically remove the fog.

The new method showed impressive results in tests that involved taking aerial photos of cities in smog as well as drone footage in foggy conditions. The results of the research can be used to improve the technology of autonomous cars and drones, which may depend on image clarity for its operation.

The neural networks used in the new method have been trained on real images taken from the internet so they can reproduce real-time images quite accurately. This allows you to automatically defog images and improve image clarity and contrast.

The results of the study show the high efficiency of the new method compared to other techniques used to improve the clarity of images in conditions of limited air transparency. This method can be of great importance in fields such as unmanned engineering, video surveillance, meteorological research, and other fields where high image quality is required under limited visibility conditions.

The researchers plan further research to improve the method and expand its application in various fields. For example, they will try to develop custom algorithms to use the method in real time with minimal delay. Overall, the new method shows great potential for improving the quality of images in conditions of limited air transparency, which could be important to many different fields of technology and science.