five

SAR Port Functional Zone Semantic Segmentation Dataset

收藏
DataCite Commons2026-04-20 更新2026-05-04 收录
下载链接:
https://data.mendeley.com/datasets/t2f8n66c87
下载链接
链接失效反馈
官方服务:
资源简介:
This dataset is constructed based on C-band and X-band SAR imagery covering multiple ports worldwide, with annotations for water areas, bulk cargo areas, container areas, tank areas and other land areas respectively. It also includes key port facilities such as tower cranes, bulk cargo storage yards, container storage yards, oil tanks, grain tanks and approach bridges. It has been used to prove the effectiveness in introducing prior knowledge as facility masks to deep learning network to precisely segment SAR port functional zones (water areas, bulk cargo areas, container areas, tank areas and other land areas). The attempt to parse these areas as a complete space is considered as an important way to further understand port functioning and benefit downstream research. The dataset comprises 5,267 sub-meter resolution images with a spatial size of 1024×1024. All samples are cropped from authentic port imagery, with a total pixel count of 5.52×10⁹. The pixel proportions of different classes exhibit significant imbalance: water area 39.91%, bulk cargo zone 3.35%, container yard 4.03%, tank farm 11.76%, and other regions 40.95%, which enables a thorough validation of the model’s generalization ability and robustness. The dataset file contains 5 files: slice_img: all SAR raw slices slice_func: functional zone masks slice_fac: facility masks We suggest using python code to attain colorful images. (We provide a dp_process.py. It is a python code to form the dataset.) # 4. color ditribution FUNC_COLOR_MAP = { 0: (255, 0, 0), # water area→blue 1: (0, 0, 255), # cargo area→red 2: (0, 255, 0), # container area→green 3: (255, 255, 0), # tank area→yellow 4: (128, 128, 128) # others→grey } FAC_COLOR_MAP = { 0: (0, 0, 0), # Background — Pure Black 1: (255, 0, 255), # Crane — Magenta / Bright Purple 2: (255, 200, 0), # Oil tank — Golden Yellow 3: (0, 255, 255), # Approach bridge — Cyan / Aqua Blue 4: (0, 180, 0), # Container yard — Dark Green 5: (255, 60, 0) # Cargo yard — Orange Red / Scarlet Red }
提供机构:
Mendeley Data
创建时间:
2026-04-15
5,000+
优质数据集
54 个
任务类型
进入经典数据集
二维码
社区交流群

面向社区/商业的数据集话题

二维码
科研交流群

面向高校/科研机构的开源数据集话题

数据驱动未来

携手共赢发展

商业合作