Convolutional Neural Networks Based Image Classification for Himawari-8 Stationary Satellite Imagery

Cloud plays an extremely important role in the atmosphere, which directly influences the radiation balance and indicates the potential weather and climate change. It also builds up the strength of locally thermal and dynamic processes. A novel convolutional neural networks (CNNs) approach, namely Satellites Model (SatMod), is introduced to satellite imagery classification, which performs well under multiple contrails conditions. We take contrails into account in the satellite imagery classification, which leads to satellite imagery classification more challenging than existing satellite imagery database, as clear understanding of contrails would facilitate the study of the effects of contrails on global warming. A novel dataset based on Himawari-8 stationary satellite imagery (HSSI) is proposed to represent 5 different scenes. Extensive experiments and evaluation indicate that the proposed SatMod achieves a good performance on HSSI database.

[1]  Matti Pietikäinen,et al.  Gray Scale and Rotation Invariant Texture Classification with Local Binary Patterns , 2000, ECCV.

[2]  Jinglin Zhang,et al.  Verification for Different Contrail Parameterizations Based on Integrated Satellite Observation and ECMWF Reanalysis Data , 2017 .

[3]  H. Hetzheim,et al.  Automated contrail and cirrus detection using stochastic properties , 2007 .

[4]  Ronald M. Welch,et al.  A neural network approach to cloud classification , 1990 .

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

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

[7]  P. Minnis,et al.  Estimation of 2006 Northern Hemisphere contrail coverage using MODIS data , 2013 .

[8]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[9]  Ulrich Schumann,et al.  Regional radiative forcing by line-shaped contrails derived from satellite data , 2002 .

[10]  Yoshua Bengio,et al.  Deep Sparse Rectifier Neural Networks , 2011, AISTATS.

[11]  Kenneth A. Buch,et al.  Cloud classification using whole-sky imager data , 1995 .

[12]  Zhiguo Cao,et al.  Cloud Classification of Ground-Based Images Using Texture–Structure Features , 2014 .

[13]  Stefan Winkler,et al.  Categorization of cloud image patches using an improved texton-based approach , 2015, 2015 IEEE International Conference on Image Processing (ICIP).

[14]  Sergey Ioffe,et al.  Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift , 2015, ICML.

[15]  B. Kärcher,et al.  Global radiative forcing from contrail cirrus , 2011 .

[16]  Matti Pietikäinen,et al.  Multiresolution Gray-Scale and Rotation Invariant Texture Classification with Local Binary Patterns , 2002, IEEE Trans. Pattern Anal. Mach. Intell..

[17]  Shuang Liu,et al.  Learning group patterns for ground-based cloud classification in wireless sensor networks , 2016, EURASIP J. Wirel. Commun. Netw..

[18]  Chun-tong Liu,et al.  Modeling and analyzing interference signal in a complex electromagnetic environment , 2016, EURASIP J. Wirel. Commun. Netw..

[19]  A. Heinle,et al.  Automatic cloud classification of whole sky images , 2010 .

[20]  Li Fei-Fei,et al.  ImageNet: A large-scale hierarchical image database , 2009, CVPR.