用于农田分割的大规模图像文本数据集基准

理解和掌握农田的时空特征对于准确的农田分割至关重要。传统的深度学习范式仅依赖于标注数据，在表示农田元素与周围环境之间的空间关系方面存在局限性。它难以有效建模农田的动态时空演变和空间异质性。语言作为一种结构化的知识载体，能够明确表达农田的时空特征，如其形状、分布以及周围环境信息。因此，...

Understanding and mastering the spatiotemporal characteristics of farmland is crucial for accurate farmland segmentation. Traditional deep learning paradigms only rely on labeled data, and have limitations in representing the spatial relationships between farmland elements and their surrounding environments. They struggle to effectively model the dynamic spatiotemporal evolution and spatial heterogeneity of farmland. Language, as a structured knowledge carrier, can explicitly express the spatiotemporal characteristics of farmland, such as its shape, distribution, and surrounding environmental information. Therefore, ...