基于无人机高精度DEM数据的梯田自动提取算法

提出一种基于无人机高精度DEM数据的梯田自动提取算法. 首先,基于无人机航测技术获取1m分辨率的DEM数据,在此基础上计算地面坡度. 此时,梯田田坎处于坡度的极大值区. 其次,采用坡面流水模拟算法实现对该极值区的提取. 最后,对提取结果进行掩膜滤波,消除非梯田区域,得到最终梯田提取结果. 以陕西省长武县王东沟流域为实验样区,进行梯田提取实验,并将实验提取结果与梯田目视解译结果及光照晕渲模拟方法提取结果进行对比. 结果显示,相比目视真值,该方法提取梯田的准确率为90.67%,具有较高的提取精度; 相比光照晕渲模拟方法,该方法可提取出更为完整和连续的梯田结果. 该梯田快速提取方法及结果可为水土保持、农业发展、生态环境治理等研究提供依据.

An Automatic Extraction Algorithm for Terraced FieldsBased on UAV High Precision DEM Data

This paper proposes an automatic extraction algorithm for terraced fields based on DEM slope characteristics. Firstly,one-meter resolution DEM is obtained with the basis of UAV aerial survey technology,and slope is calculated on this basis. Meanwhile,terrace line is located in the maximum area of slope digital terrain model. Secondly,the slope flow simulation algorithm is used to extract the extreme value region. Finally,the extraction results are masked to eliminate the non-terraced areas,and the final terraced extraction results are obtained. Taking Wangdonggou watershed in Changwu County of Shaanxi Province as an experimental sample area,the experimental results are compared with the results of terrace extraction,which are labeled by visual interpretation based on remote sensing image,or extracted from illumination model of DEM shading. The comparison results show that the accuracy of this method is 90.67%. Compared with true value of visual interpretation,it has a higher extraction accuracy. Compared with the results of illumination model of DEM shading,this method can provide more complete and continuous result of terrace. The rapid extraction method and results of the terrace can provide a basis for the research of soil and water conservation,agricultural development and ecological environment control.