基于卫星三线阵影像生成DEM和DOM研究

Three Linear Array Based on Satellite Images to Generate DEM and DOM

研究由三线阵CCD影像生成DEM和DOM。首先采用Moravec算子提取特征点,然后利用物方匹配获取的前视、中视和后视影像上同名像点,获取地面点的三维坐标,最后由移动曲面平滑法生成DEM并利用生成的DEM由反解法生成DOM。研究结果的精度评定中,一方面采用人工目视分析法研究了兴趣点提取和同名像点匹配的合理性;另一方面通过误差理论和人工量测,获取RPC前方交会、生成DEM和DOM的精度。其中,RPC前方交会中经度、纬度和高程的平均精度分别为0.629 4 m、0.487 1 m和1.213 2 m;生成DEM的平均精度为1.051 m;生成的DOM在经度和纬度方向平均精度分别为3.413 4 m和3.910 2 m。相关精度符合1:5万制图标准,说明得到的DEM和DOM有实质的地理空间信息意义。

This study mainly focuses on generating digital products of DEM and DOM with three-line array satellite CCD images. First, Moravec algorithm is used to extract feature points. Then through the means of VLL matching, it gets the corresponding points in front view, nadir view and back view images. With the corresponding points, the three-dimensional coordinates of ground points can be obtained. After that, DEM is generated using Mobile Surface Smoothing and then DOM is created by the means of Anti-solution using the DEM. To estimate the accuracy of the results, manual visual analysis is used to estimate the interest point extraction and corresponding point matching. While with help of the Error Theory and manual measurements, the accuracy of RPC forward intersection, DEM and DOM can be obtained. In RPC forward intersection, the average accuracy of longitude, latitude and elevation is 0.629 4 m, 0.487 1 m and 1.213 2 m. The average accuracy of the DEM is 1.051 0 m. And the DOM in latitude and longitude is 3.413 4 m and 3.910 2 m. The relevant accuracy of results meets 1:50 000 mapping standard, indicating the DEM and DOM have substantial geospatial information significance.