Jilin Dual-Temporal Farmland Shelterbelt Dataset

The Jilin Dual-Temporal Farmland Shelterbelt Dataset (JL-DTFS) was developed to support fine-scale mapping and structural analysis of farmland shelterbelts in complex agricultural landscapes. The dataset integrates high-resolution multispectral imagery acquired during complementary phenological stages with manually annotated shelterbelt masks, capturing seasonal variations in vegetation characteristics and the spatial complexity of narrow linear structures. Particular attention was given to shelterbelt boundaries, structural continuity, and junction connectivity to facilitate the identification of fragmented and interconnected shelterbelt networks. By incorporating complementary temporal observations and spatially independent samples, JL-DTFS provides a valuable resource for investigating multi-temporal feature fusion, semantic segmentation, and topology-aware remote sensing methods. The dataset also supports the evaluation of segmentation accuracy and structural consistency, contributing to the development of reliable shelterbelt mapping approaches and subsequent landscape-connectivity assessments.