Visibility Detection Method And System Design Based On Digital Photography

Atmospheric visibility has a direct relationship with the land transport, aviation,marine, aerial photography, astronomical observations, and military actions. Lowvisibility will have many adverse effects and even cause great harm on ferry transport,civil aviation, highway, power supply and the daily life of the people. Due to thedeterioration of the environment, low visibility caused by the haze appeared more thanbefore, the requirement of visibility detection is more urgent. But currently usedhazemeters have many disadvantages, such as expensive, complex to install, smallsample space, difficult to calibration, etc.. Therefore, it’s necessary to design a lower andconvenient visibility detection system. A visibility detection method based on digitalphotography is studied. And a system based on that method is designed.In this paper, a method that based on Dark Chanel Prior is used to detect daytimevisibility. According to the Dark Channel Prior, we estimate the transmission from adigital image. And we can detect the visibility from the transmission basis on daytimevisibility detection principle. We optimize the transmission through Guided Filtering. Inorder to estimate the atmospheric light in the image, we combine an automaticallyatmospheric light estimation method with HCI-based (human-computer interaction)method. And a visibility detection system is designed based on IP Camera and PersonalComputer, the software is developed on C#.NET platform and MATLAB.Experiment results show that the method is reliable and the error ranges of thedetection results are within20%in all experiments.