论文信息 - Convolutional Neural Network Inception-v3: A Machine Learning Approach for Leveling Short-Range Rainfall Forecast Model from Satellite Image

Convolutional Neural Network Inception-v3: A Machine Learning Approach for Leveling Short-Range Rainfall Forecast Model from Satellite Image

In this paper, we investigated the capability of artificial intelligence using one of the advanced convolutional neural networks (CNN) called inception-v3 model to forecast leveling of daily rainfall. The input of this model were the satellite images from areas in Thailand and neighboring areas. The output of the model was leveling of daily rainfall prediction. The training from scratch of inception-v3 model with the areas of satellite image got highest accuracy from our previous work. We improved the model in such way that the model outperformed our previous model by giving the information about the movement of the cloud, air mass and weather with a batch satellite. Klong Yai rain station in Rayong province of eastern region of Thailand was selected as our case study. The dataset contains satellite images of July, August, September and October 2017. 75% of the data was used for training and 25% was used as testing data. The result of forecasting reveal that the models were able to predict leveling of 1 day ahead rainfall, 2 days ahead rainfall and 3 days ahead rainfall successfully.

[1] Geoffrey E. Hinton,et al. ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[2] Dit-Yan Yeung,et al. Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting , 2015, NIPS.

[3] F ROSENBLATT,et al. The perceptron: a probabilistic model for information storage and organization in the brain. , 1958, Psychological review.

[4] Phisan Kaewprapha,et al. Daily rainfall forecast model from satellite image using Convolution neural network , 2018, 2018 International Conference on Information Technology (InCIT).

[5] Sergey Ioffe,et al. Rethinking the Inception Architecture for Computer Vision , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[6] Lior Wolf,et al. A Dynamic Convolutional Layer for short rangeweather prediction , 2015, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[7] Sergey Ioffe,et al. Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning , 2016, AAAI.