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

Accurate crop mapping provides important and timely information for decision support on the estimation of crop production at large scale. Most existing crop-specific cover products based on remote sensing data and machine learning algorithms cannot serve large agriculture production areas as a result of poor model transfer capabilities. Developing a generalizable crop classification model for spatial transfer across regions is greatly needed. A deep learning approach, named DeepCropMapping (DCM), has been developed based on long short-term memory structure with attention mechanisms through integrating multi-temporal and multi-spectral remote sensing data for large-scale dynamic corn and soybean mapping. Transformer, Random Forest (RF), and Multilayer Perceptron (MLP) models were built for comparison. The results of the in-season classification experiment indicated the DCM model captured critical information from key growth phases and achieved higher accuracy than other models after the beginning of July. By monitoring the classification confidence in each time step, the results showed that the increased length of seasonal remote sensing time series would reduce the classification uncertainty in all sites. This study provided a viable option toward large-scale dynamic crop mapping through the integration of deep learning and remote sensing time series.

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

Master candidate Xu Jinfan in Prof. Lin Tao’s team was the first author of the paper. Corresponding authors of the paper were Prof. Lin Tao, 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 Central South University. This work was funded by National Natural Science Foundation of China and Zhejiang University.

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