Development of visibility expectation system based on machine learning

Abstract

Visibility impairment is maximum definitely defined because the formation of haze that obscures the clarity, shade, texture, and form of what’s visible through the atmosphere. It’s far a complex phenomenon inspired via some of the emissions and air pollutants and tormented by some of the herbal factors which include temperature, humidity, meteorology, time and sunlight. The aim of the research is that to estimate weather visibility using machine learning techniques. We use images taken from CCTV cameras as inputs and deep convolutional neural network model to predict results. We implemented Java based GUI application that can flexibly operate all operations in real-time. Users are also able to use a specially built web page to estimate visibility that a built-in machine learning (ML) model gives an opportunity to the user to get results. In this paper, we will detail explain regarding an architecture of the ML model, System Structure, and other essential details. © Springer Nature Switzerland AG 2018.