基于高分影像的黄土高原梯田提取技术研究

发布时间：2018-05-22 20:57

本文选题：遥感影像 + 黄土高原　；参考：《西北农林科技大学》2017年硕士论文

【摘要】：梯田是治理坡面水土流失的一项行之有效的水土保持工程措施,具有明显的保水、保土与保肥效用,在我国黄土高原地区的水土流失防治工作中发挥着不可替代的作用。梯田调查是水土保持动态监测和评价工作中不可或缺的一环,准确快速的梯田信息提取为估算区域尺度水土流失量及厘定梯田在水沙减少中的作用提供理论依据和技术支持。目前黄土高原梯田调查方法较为落后,采用调查统计法确定梯田面积,其速度慢且精度低,不能满足现阶段的工作需求,亟需探索出更为快速高效的梯田信息调查方法。随着高分辨率遥感影像的普及,其作为信息源开展黄土高原梯田信息监测不仅十分重要,而且十分迫切。论文以黄土高原丘陵沟壑区典型小流域燕沟为研究区,以高分一号遥感影像为数据源,基于面向对象分类技术对梯田信息的提取进行了探索。通过影像分割构建影像对象,在总结分析影像对象的光谱、纹理和空间等特征的基础上,根据多次试验确定最优分割合并尺度参数,构建了黄土高原梯田信息提取流程,从而实现了计算机自动为主,人机交互为辅的梯田信息提取。本研究的主要结果如下:(1)构建合理完整的梯田提取技术流程,包括辐射定标、大气校正、正射校正、图像融合、面向对象特征提取、保存对象层,提取出燕沟流域中的全部梯田,确定流域梯田分布与面积。(2)采用面向对象的多尺度图像分割与合并技术进行梯田信息提取时,使用ENVI面向对象空间特征提取模块,根据预览窗口,经过多次试验反复对比分析,选取合适的尺度因子,对研究区影像进行多尺度分割,最终确定了适于梯田提取的最优分割与合并阈值分别为36%与85%。(3)选取有代表性梯田用南方灵锐S86差分GPS进行实地定点面积测量,将实地测量面积值与影像中提取面积值进行对比,确定梯田提取精度较高,从而证明此提取方法流程是可行的。总结了高分一号卫星梯田信息提取的最优尺度和梯田信息提取方法及技术流程,为水土保持工程措施梯田的监测提供技术参考。
[Abstract]:Terrace is an effective water and soil conservation project to control soil erosion on the slope. It has obvious water conservation, soil conservation and fertilizer conservation. It plays an irreplaceable role in the prevention and control of soil and water loss in the Loess Plateau area of China. The extraction of fast terrace information provides theoretical basis and technical support for estimating the water and soil loss of regional scale and determining the effect of terracing in water and sediment reduction. At present, the investigation method of terraced field in the Loess Plateau is relatively backward. The area of terraced field is determined by the investigation and statistics method, which is slow and low in precision, and it can not meet the needs of the present stage. With the popularization of high resolution remote sensing images, with the popularization of high resolution remote sensing images, it is very important and urgent to carry out the information monitoring of the terrace in the Loess Plateau as a source of information. On the basis of analyzing and analyzing the characteristics of the spectral, texture and space of the image objects, based on the analysis of the characteristics of the spectral, texture and space of the image objects, the information extraction process of the terrace in the Yellow Soil Plateau is constructed, and the calculation of the information extraction process of the terraced terrace in the Yellow Soil Plateau is constructed. The main results of this study are as follows: (1) constructing a reasonable and complete technological process for the extraction of terraced fields, including radiation calibration, atmospheric correction, orthophoto correction, image fusion, object feature extraction, and preservation of object layers, extracting all terraces in the valley of Yan Gou, and determining terraced fields in the basin. Cloth and area. (2) using object-oriented multi-scale image segmentation and merging technology to extract information from terraced fields, using ENVI object oriented space feature extraction module, according to preview window, through repeated experiments and repeated comparison analysis, select the appropriate scale factor, multiscale segmentation of the image of the study area, and finally determine the suitable terrace. The optimal segmentation and merging thresholds were 36% and 85%. (3) respectively. The field site area measurement was measured by the southern Ling sharp S86 difference GPS of the representative terraced field. The field measurement area was compared with the value of the extracted area in the image, and the extraction precision of the terraced field was higher. Thus it was proved that the extraction method was feasible. The optimal scale of information extraction from terraced fields and information extraction methods and technological process of terraced fields are provided to provide technical reference for monitoring terraced fields in soil and water conservation projects.
【学位授予单位】：西北农林科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：S157.31;S127

【相似文献】