Development of visibility expectation system based on machine learning
- Authors
- Palvanov, A.; Giyenko, A.; Cho, Y.I.
- Issue Date
- 2018
- Publisher
- Springer Verlag
- Keywords
- CNN; GUI; Machine learning; Visibility
- Citation
- Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), v.11127 LNCS, pp.140 - 153
- Journal Title
- Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
- Volume
- 11127 LNCS
- Start Page
- 140
- End Page
- 153
- URI
- https://scholarworks.bwise.kr/gachon/handle/2020.sw.gachon/4314
- DOI
- 10.1007/978-3-319-99954-8_13
- ISSN
- 0302-9743
- 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.
- Files in This Item
- There are no files associated with this item.
- Appears in
Collections - IT융합대학 > 컴퓨터공학과 > 1. Journal Articles
![qrcode](https://api.qrserver.com/v1/create-qr-code/?size=55x55&data=https://scholarworks.bwise.kr/gachon/handle/2020.sw.gachon/4314)
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.