Mars-Bench: mb-mars_seg_msl

The mb-mars_seg_msl dataset provides multi-class terrain segmentation from the Mars Science Laboratory (MSL) mission, captured by the Curiosity rover’s Mastcam instrument. It includes high-resolution visual imagery of diverse Martian terrains, supporting studies in surface composition, navigation, and rover path planning. This dataset is annotated with six terrain-related classes: Bedrock, Gravel/Sand/Soil, Rock, Shadow, Sky/Distant Mountains, and Track; representing the dominant visual elements in rover imagery. To explore all Mars-Bench resources, visit our project website: https://mars-bench.github.io/mb-mars_seg_msl 数据集收录了源自火星科学实验室（Mars Science Laboratory，MSL）任务的多类别地形分割数据，这些数据由好奇号漫游车（Curiosity rover）的桅杆相机（Mastcam）采集获取。该数据集包含多样火星地形的高分辨率视觉影像，可用于地表成分分析、漫游车导航与路径规划等相关研究。数据集共标注了6类地形相关类别：基岩（Bedrock）、砾石/沙/土壤（Gravel/Sand/Soil）、岩石（Rock）、阴影（Shadow）、天空/远山（Sky/Distant Mountains）以及车辙（Track），覆盖了漫游车影像中的典型视觉元素。 如需浏览全部Mars-Bench相关资源，请访问项目官网：https://mars-bench.github.io/