Outdoor large-scale scene point cloud segmentation method based on A-EdgeConv

The invention belongs to the technical field of point cloud image processing, and particularly relates to an A-EdgeConv-based outdoor large-scale scene point cloud segmentation method, which combines local geometric information and a graph cut algorithm to realize super-point acquisition, and adopts a mode of combining a local adjacency graph with an attention mechanism to more accurately extract super-point features, so that the accuracy of point cloud segmentation is improved. Through the GRU recurrent neural network, super-point and super-edge feature aggregation is realized, the point cloud segmentation speed in a large scene is greatly improved, feature expressive force is improved due to combination of a local adjacency graph and an attention mechanism, effective information and a point cloud local structure are concerned, segmentation precision is improved, and optimization complexity is reduced. 本发明属于点云图像处理技术领域，具体涉及一种基于A-EdgeConv的室外大场景点云分割方法，结合局部几何信息与图割算法实现超点的获取，并采用局部邻接图结合注意力机制的方式更加准确的提取超点特征，通过GRU循环神经