Deep learning for Amazon satellite image analysis

Machine learning can be the key to saving the world from losing football field-sized forest areas each second. As deforestation in the Amazon basin causes devastating effects both on the ecosystem and the environment, there is urgent need to better understand and manage its changing landscape. A competition was recently conducted to develop algorithms to analyze satellite images of the Amazon. Successful algorithms will be able to detect subtle features in different image scenes, giving us the crucial data needed to be able to manage deforestation and its consequences more effectively. This paper presents our entry to the competition, the results obtained, and possible improvements to the algorithm.

[1]  L. Losik.  Using satellites to predict earthquakes, volcano eruptions, identify and track tsunamis from space , 2012, 2012 IEEE Aerospace Conference.

[2]  Steve Ankuo Chien,et al.  Remote Sensing of Natural Disasters , 2013 .

[3]  G. Popkin.  Satellite alerts track deforestation in real time , 2016, Nature.

[4]  S. Frühwirth-Schnatter.  Data Augmentation and Dynamic Linear Models , 1994 .

[5]  Sebastiano B. Serpico,et al.  Remote Sensing of Natural Disasters , 2012, Proc. IEEE.

[6]  Michael S. Bernstein,et al.  ImageNet Large Scale Visual Recognition Challenge , 2014, International Journal of Computer Vision.

[7]  Jian Sun,et al.  Object Detection Networks on Convolutional Feature Maps , 2015, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[8]  Juan José Rodríguez Diez,et al.  A weighted voting framework for classifiers ensembles , 2012, Knowledge and Information Systems.

[9]  Jiebo Luo,et al.  Learning multi-label scene classification , 2004, Pattern Recognit..

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

[11]  Alberto Refice,et al.  SAR and InSAR for Flood Monitoring: Examples With COSMO-SkyMed Data , 2014, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.

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

[13]  Andrew Zisserman,et al.  Very Deep Convolutional Networks for Large-Scale Image Recognition , 2014, ICLR.

[14]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[15]  Masaki Aono,et al.  Residual Network with Delayed Max Pooling for Very Large Scale Plant Identification , 2017, CLEF.

[16]  Ranga B. Myneni,et al.  Monitoring Rainforest Dynamics in the Amazon with MODIS Land Products , 2006, 2006 IEEE International Symposium on Geoscience and Remote Sensing.

[17]  Forrest N. Iandola,et al.  DenseNet: Implementing Efficient ConvNet Descriptor Pyramids , 2014, ArXiv.

[18]  Bernhard Rabus,et al.  Sensitivity of Barnes Ice Cap, Baffin Island, Canada, to climate state and internal dynamics , 2016 .

[19]  Stan Szpakowicz,et al.  Beyond Accuracy, F-Score and ROC: A Family of Discriminant Measures for Performance Evaluation , 2006, Australian Conference on Artificial Intelligence.

[20]  François Chollet,et al.  Xception: Deep Learning with Depthwise Separable Convolutions , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[21]  Oskar Levander,et al.  Autonomous ships on the high seas , 2017, IEEE Spectrum.