基于远距离场景影像的露天矿边坡三维重建技术

Technology of Three-dimensional Reconstruction of Opencast Slope based on Distant Scenes Image

露天矿边坡受采矿工艺及自然条件的影响,其几何形状复杂,传统测量方法难以全面的进行描述。将基于影像的三维自动重建技术引入计算机视觉领域,建立一种适用于远距离场景的透视相机模型,利用仿射相机模型的鲁棒自标定算法,通过非量测数码拍摄露天矿边坡相片即可全自动构建出边坡的三维模型。依据上述远距离场景影像获取和处理方法,对鞍千矿业公司某一边坡进行了现场测试,获得了不同数量照片条件下边坡三维模型,说明该种露天矿边坡三维建模技术是可行的,并且原始图片数量相对较少的情况下也可获得较高精度的模型,有效地降低露天矿边坡三维模型数据采集的工作量、减小三维建模成本、提高露天矿测绘效率,在露天矿三维地形、滑坡灾害分析等方面具有重要的意义。

Due to the influence by the mining process and the natural conditions,the geometric shape of open-pit slope is complex and it is difficult to describe completely by the traditional measurement method.The technology of three-dimensional automatic reconstruction based on imagine is introduced to the computer vision field,and then,a perspective camera model for the distant scenes imagine is built.With the use of robust self-calibration algorithm,a three-dimensional model of slope can be automatically built through the non-metric digital open-pit mine slope photos.According to the acquisition and processing of the above distant-scene image,a field test on a slope which belongs to Anqian Mining Company was carried out.A three-dimensional slope model was acquired under different numbers of the photos,which means that the three-dimensional modeling technique of open pit slope is feasible.Moreover,higher precision model could be acquired under the condition of less original pictures.So it can effectively reduce the work of the data collection and the cost of three-dimensional modeling and improve the efficiency of surveying and mapping open-pit mine.It has important significance in the field of three-dimensional terrain and landslide hazard analysis in open pit mine.