文献类型： 外文期刊

作者： Cai, Zhiwen ¹ ; Xu, Baodong ¹ ; Yu, Qiangyi ² ; Zhang, Xinyu ³ ; Yang, Jingya ² ; Wei, Haodong ³ ; Li, Shiqi ¹ ; Song, Qian ² ; Xiong, Hang ⁴ ; Wu, Hao ⁵ ; Wu, Wenbin ² ; Shi, Zhihua ¹ ; Hu, Qiong ⁵ ;

作者机构： 1.Huazhong Agr Univ, Coll Resources & Environm, Wuhan 430070, Peoples R China

2.Chinese Acad Agr Sci, Inst Agr Resources & Reg Planning, State Key Lab Efficient Utilizat Arid & Semiarid A, Beijing 100081, Peoples R China

3.Huazhong Agr Univ, Macro Agr Res Inst, Coll Plant Sci & Technol, Wuhan 430070, Peoples R China

4.Huazhong Agr Univ, Coll Econ & Management, Wuhan 430070, Peoples R China

5.Cent China Normal Univ, Coll Urban & Environm Sci, Key Lab Geog Proc Anal & Simulat Hubei Prov, Wuhan 430079, Peoples R China

关键词： Crop type mapping; Limited sparse point annotations; Weakly supervised semantic segmentation; multi-temporal Sentinel-2 images

期刊名称：ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING （ 影响因子：12.2； 五年影响因子：13.7 ）

ISSN： 0924-2716

年卷期： 2024 年 218 卷

页码：

收录情况： SCI

摘要： Accurate and timely information on the spatial distribution and areas of crop types is critical for yield estimation, agricultural management, and sustainable development. However, traditional crop classification methods often struggle to identify various crop types effectively due to their intricate spatiotemporal patterns and high training data demands. To address this challenge, we developed a Structure-aware Label eXpansion segmentation Network (StructLabX-Net) for diverse crop type mapping using limited point-annotated samples. StructLabX-Net features a backbone U-TempoNet, which combines CNNs and LSTM to explore intricate spatiotemporal patterns. It also incorporates multi-task weak supervision heads for edge detection and pseudo-label expansion, adding crucial structure and contextual insights. We tested the StructLabX-Net across three distinct regions in China, assessing over 10 crop types and comparing its performance against five popular classifiers based on multitemporal Sentinel-2 images. The results showed that StructLabX-Net significantly outperformed RF, SVM, DeepCropMapping, Transformer, and patch-based CNN in identifying various crop types across three regions with sparse training samples. It achieved the highest overall accuracy and mean F1-score: 91.0% and 89.1% in Jianghan Plain, 91.5% and 90.7% in Songnen Plain, as well as 91.0% and 90.8% in Sanjiang Plain. StructLabXNet demonstrated a particular advantage for those "hard types" characterized by limited samples and complex phenological features. Furthermore, ablation experiments highlight the crucial role of the "edge" head in guiding the model to accurately differentiate between various crop types with clearer class boundaries, and the "expansion" head in refining the understanding of target crops by providing extra details in pseudo-labels. Meanwhile, combining our backbone U-TempoNet with multi-task weak supervision heads exhibited superior results of crop type mapping than those derived by other segmentation models. Overall, StructLabX-Net maximizes the utilization of limited sparse samples from field surveys, offering a simple, cost-effective, and robust solution for accurately mapping various crop types at large scales. The code will be publicly available at htt ps://github.com/BruceKai/StructLabX-Net.