Method and system for generating annotation data set in entity scene

The invention discloses a method and system for generating an annotation data set by an entity scene, and the method comprises the steps: 1, obtaining an entity scene which comprises a plurality of annotation targets, the annotation targets are features of the entity scene, and each annotation target is provided with a boundary; 2, sensing, labeling and extracting each labeling target in the entity scene by using a full-dimensional labeling method; 3, de-duplicating the extracted annotation targets, and splicing to generate a scene annotation data set; and 4, applying the scene annotation data set to a data visualization tool for scene simulation. According to the method, the working efficiency is greatly improved, repeated data operation is completely avoided, and a large number of resources are saved. 本申请公开了一种实体场景生成标注数据集的方法和系统，所述方法包括：步骤1：获取实体场景，所述实体场景包含若干标注目标，所述标注目标为实体场景的特征，每一标注目标均有边界；步骤2：使用全维度标注法对实体场景中的每一标注目标进行感知、标注和提取；步骤3：对提取到的标注目标去重并拼接生成场景标注数据集；步骤4：将所述场景标注数据集应用在数据可视化工具进行场景仿真。本发明极大提高了工作效率，完全避免了数据重复操作，节省了大