Improving Convective Cloud Classification with Deep Learning: The CC-Unet Model

Abstract

Analyzing and mitigating natural disasters can be a challenging task, which is why the field of computer science, specifically artificial intelligence (AI) is necessary to aid in the complexity of disaster management. AI provides the tools and analytical models to help solve the intricacies of handling natural disasters. Convective clouds, closely related to rain and can lead to large-scale, prolonged hydrometeorological disasters, are a crucial component to consider. To improve the classification of these clouds, a predictive-analytical model based on deep learning, called the CC-Unet model, was developed. This model utilizes a U-Net architecture and is trained using a dataset of convective cloud images. The researchers used satellite image data from the Himawari 8 satellite collected in May and October 2021. The images were pre-processed and verified using observational data. The model was tested using a random train-test split method, showing that the CC-Unet model had a higher accuracy of 97.29% compared to the U-Net model, which had an accuracy of 94.17%. Additionally, the significance test using the Wilcoxon method showed that the CC-Unet model had significantly different performance results from the U-Net model. The ground truth image was also compared with the predicted image, showing a low root mean square error value of 0.0218, indicating a high level of similarity between the two. Overall, this research demonstrates the potential of AI and deep learning in classifying convective clouds to aid in natural disaster management.

Methodology for Severe Convective Cloud Identification Using Lightweight Neural Network Model Ensembling

“…Cloud segmentation based on deep learning is a long-standing and ongoing area of research in the field of remote sensing. This includes general cloud segmentation without differentiating cloud types [5,11,14,[17][18][19][20][21][22], as well as segmentation of severe convective clouds [2,4,6,8,15,[23][24][25][26][27], which continue to be actively studied, underscoring the enduring significance of this problem. In practical applications, particularly where real-time processing of vast datasets is required, lightweight neural networks have garnered attention due to their lower computational demands and rapid processing capabilities.…”

“…By aggregating predictions from multiple models, ensemble methods can effectively reduce the error rates that might occur in individual models, thereby achieving higher accuracy [3,42,43]. Additionally, ensemble methods enhance the predictive capability of models on unseen samples, which is particularly vital under the fluctuating conditions of meteorological phenomena [24,25,44]. As lightweight neural network technologies [45] continue to advance, their integration into ensemble frameworks for recognizing severe convective clouds offers significant research value and practical prospects.…”

“…For light, medium, or large neural networks, researchers have made many meaningful and significant improvements to make them more adaptable to cloud recognition tasks [20,24,42,[46][47][48][49][50] (including the identification of severe convective clouds). However, beyond these improvements, we can also consider how to use model ensemble techniques to better integrate and utilize existing lightweight neural networks, enabling them to exhibit capabilities and computational efficiency far beyond a single large neural network.…”

“…Given the ample excellent research in the field [2,4,[6][7][8]15,[23][24][25][26][27]59,60], we do not propose a specific neural network. While developing new networks has its merits, optimizing the use of existing networks is equally valuable [61].…”