Time Gated Convolutional Neural Networks for Crop Classification
This work addresses crop classification for agricultural monitoring, presenting a domain-specific incremental improvement.
The paper tackled crop classification using earth observation time series by proposing a Time Gated Convolutional Neural Network (TGCNN) that incorporates temporal information and gating mechanisms, achieving an F1 score of 0.973, AUC ROC of 0.977, and IoU of 0.948, and outperforming benchmarks in tasks across Kenya, Brazil, and Togo.
This paper presented a state-of-the-art framework, Time Gated Convolutional Neural Network (TGCNN) that takes advantage of temporal information and gating mechanisms for the crop classification problem. Besides, several vegetation indices were constructed to expand dimensions of input data to take advantage of spectral information. Both spatial (channel-wise) and temporal (step-wise) correlation are considered in TGCNN. Specifically, our preliminary analysis indicates that step-wise information is of greater importance in this data set. Lastly, the gating mechanism helps capture high-order relationship. Our TGCNN solution achieves $0.973$ F1 score, $0.977$ AUC ROC and $0.948$ IoU, respectively. In addition, it outperforms three other benchmarks in different local tasks (Kenya, Brazil and Togo). Overall, our experiments demonstrate that TGCNN is advantageous in this earth observation time series classification task.