On June 08, 2020, Prof. Lin Tao’s team publisheda paper entitled “DeepCropMapping: A multi-temporal deep learning approach withimproved spatial generalizability for dynamic corn and soybean mapping” in thejournal of Remote Sensing of Environment. This study built aDeepCropMapping (DCM) model based on a Long Short-Term Memory structure withattention mechanism. The DCM model improved spatial generalizability for cornand soybean mapping, and learned accumulated time-series spectral features. Thestudy provided a viable deep learning approach toward large-scale crop mapping.

Accurate crop mappingprovides important and timely information for decision support on theestimation of crop production at large scale. Most existing crop-specific coverproducts based on remote sensing data and machine learning algorithms cannotserve large agriculture production areas as a result of poor model transfercapabilities. Developing a generalizable crop classification model for spatialtransfer across regions is greatly needed. A deep learning approach, namedDeepCropMapping (DCM), has been developed based on long short-term memorystructure with attention mechanisms through integrating multi-temporal andmulti-spectral remote sensing data for large-scale dynamic corn and soybeanmapping. Transformer, Random Forest (RF), and Multilayer Perceptron (MLP)models were built for comparison. The results of the in-season classificationexperiment indicated the DCM model captured critical information from keygrowth phases and achieved higher accuracy than other models after thebeginning of July. By monitoring the classification confidence in each timestep, the results showed that the increased length of seasonal remote sensingtime series would reduce the classification uncertainty in all sites. Thisstudy provided a viable option toward large-scale dynamic crop mapping throughthe integration of deep learning and remote sensing time series.

The long abstract of a previous work of this study entitled “Efficientmulti-temporal and in-season Crop Mapping with Landsat Analysis Ready Data viaLong short-term Memory Networks”, was presented in the seminar of 2019International Conference on Machine Learning (ICML): “Climate Change: How CanAI Help?”. ICML was an annual machine learning international top conference,which was sponsored by the international machine learning society (IMLS). Thisseminar was organized by YoshuaBengi, Turing award winner in 2018, and others, aimingto discuss how to make use of artificial intelligence technology to helpsociety to adapt to climate change. Prof. Andrew Ng from Stanford University, afamous scholar in artificial intelligence field, and other researchers gave apresentation. The seminar was ICML's first workshop on climate change and receivedconsiderable attention from participants.

Master candidate Xu Jinfan in Prof. Lin Tao’s team was thefirst author of the paper. Corresponding authors of the paper were Prof. LinTao, in the College of Biosystems Engineering and Food Science, from Zhejiang University, and Prof.Li Haifeng, in the School of Geosciences and Information Physics, from CentralSouth University. This work was funded by National Natural Science Foundationof China and Zhejiang University.

Link to the article: https://www.sciencedirect.com/science/article/pii/S0034425720303163#f0005