基于SPOT5影像的1∶1万土地利用矢量数据误差研究

Error of 1∶10000 vector data for land use based on SPOT5 image

基于遥感的1∶1万土地利用更新调查矢量数据的空间匹配误差是该应用能否实现的关键。实验选取浙江省绍兴市杨汛桥镇为例,利用控制点、检测点GPS野外实测数据,以及同期IKONOS数据,采用对比法,进行了基于SPOT5遥感影像的土地利用更新调查(平原区域)矢量数据的点、面的空间匹配误差的研究,最后得出了点位误差结果:水平方向7.38m,垂直方向4.53m,点位平面中误差8.66m;以及面状地物空间匹配误差结果:绝大部分地类的面积吻合度及相对面积精度都可达95%左右。因此,基于SPOT5影像的1∶1万平原区域土地利用更新调查是完全可行的。

Land use change survey has a new operation method with the development of remote sensing technology. Compared to aerial survey or the custom method, the high resolution images of SPOT5 make it cost-efficient for remote sensing-based land use change survey at a 1 ： 10000 scale. The key point is that whether the error of the vector data based on RS can satisfy with the error limitation. In order to get the position error of the vector data for the applications, one case study was conducted in Yangxunqiao town of Shaoxing city in Zhejiang Province. The ground control points （GCPs） and check points were taken by differential GPS through field survey. At the same time, the area errory of the polygon features were evaluated by using simultaneous IKONOS images. The point position matching mean square error results were obtained and polygon feature position matching error result shows that area accuracy is mostly about 95%. Therefore, the approach to land use change survey based on 1 ： 10000 SPOT5 image in plain areas is proved feasible.