Efficient Resnet Model for Atmospheric Visibility Classification

Abstract

Atmospheric visibility is an important element of meteorological observation. With existing methods, defining image features that reflect visibility accurately and comprehensively is difficult. This paper proposes a visibility detection method based on a deep learning model termed Resnet that addresses issues caused by a lack of sufficient visibility labelled datasets. The proposed model is used for calculating the visibility of external geo-tagged images directly without relying on weather images or data that require costly sensing or customized capturing. A large collection of internet photographs is used for the current data-driven approach to learning rich scene and visibility variations. Finally, the performance of Resnet is compared over the different variant of Resnet and other deep learning models in terms of diverse performance metrics and shows the ability of proposed Resnet model in atmospheric visibility detection.