Deep Gaussian Process for Crop Yield Prediction Based on Remote Sensing Data

Agricultural monitoring, in particular in developing countries, can help prevent famine and support humanitarian efforts. A central challenge is yield estimation, which is to predict crop yields before harvesting. We introduce a scalable, accurate, and inexpensive method to predict crop yields using publicly available remote sensing data. Our approach improves existing techniques in three ways. First, we forego hand-crafted features traditionally used in the remote sensing community and propose an approach based on modern representation learning ideas. We introduce a novel dimensionality reduction technique that allows us to train a Convolutional Neural Network or Longshort Term Memory and automatically learn useful features even when labeled training data is scarce. Finally, we incorporate a Gaussian Process component to explicitly model the spatio-temporal structure of the data and further improve the accuracy. We evaluate our approach on county-level soybean production in the U.S. and show that our approach vastly outperforms competing techniques.

[1]  Hywel T. P. Williams,et al.  Using Social Media to Detect and Locate Wildfires , 2016, EcoMo@ICWSM.

[2]  Zoran Obradovic,et al.  Continuous Conditional Random Fields for Efficient Regression in Large Fully Connected Graphs , 2013, AAAI.

[3]  Theodoros Damoulas,et al.  Adaptive Spatio-Temporal Exploratory Models: Hemisphere-wide species distributions from massively crowdsourced eBird data , 2013, AAAI.

[4]  Andrew Gordon Wilson,et al.  Deep Kernel Learning , 2015, AISTATS.

[5]  Yaxing Wei,et al.  Daymet: Daily Surface Weather Data on a 1-km Grid for North America, Version 2 , 2014 .

[6]  Leo Breiman,et al.  Statistical Modeling: The Two Cultures (with comments and a rejoinder by the author) , 2001 .

[7]  Felix Dodds,et al.  The water, food, energy and climate nexus : challenges and an agenda for action , 2016 .

[8]  Stefano Ermon,et al.  Transfer Learning from Deep Features for Remote Sensing and Poverty Mapping , 2015, AAAI.

[9]  Ryosuke Shibasaki,et al.  Estimating crop yields with deep learning and remotely sensed data , 2015, 2015 IEEE International Geoscience and Remote Sensing Symposium (IGARSS).

[10]  Douglas K. Bolton,et al.  Forecasting crop yield using remotely sensed vegetation indices and crop phenology metrics , 2013 .

[11]  N. Cressie,et al.  Robust estimation of the variogram: I , 1980 .

[12]  David M. Johnson An assessment of pre- and within-season remotely sensed variables for forecasting corn and soybean yields in the United States , 2014 .

[13]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[14]  Geoffrey E. Hinton,et al.  Using Deep Belief Nets to Learn Covariance Kernels for Gaussian Processes , 2007, NIPS.

[15]  Suha Berberoglu,et al.  Crop yield prediction under soil salinity using satellite derived vegetation indices , 2016 .

[16]  Jun Yu,et al.  eBird: A Human/Computer Learning Network for Biodiversity Conservation and Research , 2012, IAAI.

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

[18]  S. McGuire,et al.  FAO, IFAD, and WFP. The State of Food Insecurity in the World 2015: Meeting the 2015 International Hunger Targets: Taking Stock of Uneven Progress. Rome: FAO, 2015. , 2015, Advances in nutrition.

[19]  Stefano Ermon,et al.  Learning Large-Scale Dynamic Discrete Choice Models of Spatio-Temporal Preferences with Application to Migratory Pastoralism in East Africa , 2015, AAAI.

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

[21]  Guigang Zhang,et al.  Deep Learning , 2016, Int. J. Semantic Comput..

[22]  Wfp,et al.  The State of Food Insecurity in the World , 2011 .

[23]  Zoubin Ghahramani,et al.  Bayesian Convolutional Neural Networks with Bernoulli Approximate Variational Inference , 2015, ArXiv.

[24]  N. Silleos,et al.  The use of multi-temporal NDVI measurements from AVHRR data for crop yield estimation and prediction , 1993 .

[25]  Carl E. Rasmussen,et al.  Gaussian processes for machine learning , 2005, Adaptive computation and machine learning.

[26]  Sang Michael Xie,et al.  Combining satellite imagery and machine learning to predict poverty , 2016, Science.

[27]  Fei-Fei Li,et al.  Large-Scale Video Classification with Convolutional Neural Networks , 2014, 2014 IEEE Conference on Computer Vision and Pattern Recognition.

[28]  Christopher Kermorvant,et al.  Dropout Improves Recurrent Neural Networks for Handwriting Recognition , 2013, 2014 14th International Conference on Frontiers in Handwriting Recognition.

[29]  G. Parisi,et al.  Statistical Field Theory , 1988 .

[30]  Un Desa Transforming our world : The 2030 Agenda for Sustainable Development , 2016 .

[31]  Zhengwei Yang,et al.  Monitoring US agriculture: the US Department of Agriculture, National Agricultural Statistics Service, Cropland Data Layer Program , 2011 .

[32]  D. Lobell,et al.  A scalable satellite-based crop yield mapper , 2015 .

[33]  Wei Chu,et al.  Preference learning with Gaussian processes , 2005, ICML.