Three-dimensional reconstruction method based on image recognition

The invention discloses a three-dimensional reconstruction method based on image recognition, which comprises the following steps of: processing an aerial picture or three-dimensional laser point cloud data acquired by an unmanned aerial vehicle by using software (intelligent algorithm) to generate a three-dimensional model, automatically generating and extracting an elevation map and a current situation map by using the three-dimensional model, and finally manually labeling and extracting a specific point location; after the unmanned aerial vehicle or the three-dimensional laser collects images and point clouds, the intelligent processing system is used for intelligently generating a three-dimensional diagram, and the intelligent processing system is used for intelligently recognizing elevation and vector diagrams. 一种基于图像识别三维重建方法，将由无人机采集的航拍图片或三维激光点云数据，利用软件(智能算法)处理生成三维模型，再利用三维模型自动生成提取高程图及现状图，最后人工对特定点位进行标注提取；无人机或三维激光采集图像和点云后，利用智能处理系统智能生成三维图，利用智能处理系统智能识别高程及矢量图。