Identify cloud patterns by both cloud images and meteorological parameters using hybrid deep learning model

Bera, Indranil (2022) Identify cloud patterns by both cloud images and meteorological parameters using hybrid deep learning model. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration manual) Download (413kB) | Preview

Abstract

The atmosphere is a complex interaction of different meteorological features. Weather prediction is one of the most prominent concerns in meteorological research. Weather forecasting can be very helpful to take preventive measures for any upcoming concerns, mitigate financial risk, life loss etc. The traditional way of forecasting is by Numerical weather prediction techniques which solve several complex differential equations related to atmospheric energy. NWP is very time-consuming and resource extensive. Deep learning algorithms become a popular way of dealing with huge weather data captured from weather stations. Cloud patterns are closely related to the weather parameters of that particular location. Several research was done on cloud pattern detection by cloud images from satellites, from ground-based camera but very less study was found by combining the cloud images and weather features. Here hybrid novel multi-modal architecture was developed using ground-based cloud images of Singapore and 9 weather features. Multi-Layer Perceptron (MLP) was implemented for numerical data and CNN-based ‘MobileNet’ pre-trained model was employed to extract features from the images. This hybrid model outperformed the previously built CNN-based pre-trained ‘VGG16’ model on the same dataset by 8%. Incorporating weather parameters in the image classification model can improve cloud classification significantly.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Metrology; Forecast; Numerical Weather Prediction (NWP); Deep Learning; CNN; Multi-Modal; Multi-Layer Perceptron (MLP); VGG16; MobileNet
Subjects:	G Geography. Anthropology. Recreation > G Geography (General) Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Tamara Malone
Date Deposited:	18 Jan 2023 17:04
Last Modified:	06 Mar 2023 16:54
URI:	https://norma.ncirl.ie/id/eprint/6085

Actions (login required)

View Item