Task scheduling method and system for logistics robot cluster

The invention provides a task scheduling method and system for a logistics robot cluster, and the method comprises the steps: obtaining the addresses of a plurality of task targets through text extraction and keyword recognition according to a logistics task package; acquiring depth image information of the surrounding environment of each robot in the logistics robot cluster, and converting the depth image information into discrete three-dimensional point cloud data; based on the three-dimensional point cloud data, a visual SLAM algorithm based on graph optimization is adopted, and a three-dimensional prior map is generated through the steps of data frame construction, key frame sequence screening, inter-frame registration execution, graph optimization and point cloud fusion optimization; semantic information labeling is carried out on the three-dimensional prior map, and construction of a hierarchical semantic map of a grid layer, a topology layer and a semantic layer is achieved; path planning is carried ou