基于ZY-3和图像融合的冲沟参数提取研究  

Extracting gully parameters using multispectral(MS) and panchromatic(PAN) fusion image of the Zi Yuan-3(ZY-3) satellite

作　　者：王涛 赵延芳[1] 何福红[1] 李鹏[1] 马江涛[1] 

机构地区：[1]鲁东大学资源与环境工程学院,山东烟台264025

出　　处：《农业现代化研究》2016年第6期1190-1197,共8页Research of Agricultural Modernization

基　　金：国家自然科学基金项目(41001160;41471223);山东省自然科学基金项目(ZR2014JL026);烟台市科技计划项目(2015ZH091)~~

摘　　要：本文以山东省栖霞市庵里水库西岸流域为研究区,基于Gram-S chimdt(GS)方法、色彩空间(Hue Saturation Value,HSV)变换和主成分分析(Principal Component Analysis,PCA)三种方法,对研究区的资源三号卫星(ZY-3)图像进行了图像融合,基于各融合图像提取冲沟参数并进行精度评价。结果表明:三种融合图像均保留了多光谱图像的光谱信息,同时具有全色图像的空间纹理细节信息;与基于原始多光谱图像解译冲沟参数相比较,采用GS光谱锐化、HSV变换及PCA变换融合图像解译冲沟参数更准确;相对检验区实测冲沟参数,基于GS光谱锐化融合图像提取的冲沟参数精度最高,解译出的冲沟沟面积和沟长的相关系数分别为0.991和0.984。是冲沟参数提取的最优图像融合方法。Selecting the west coast watershed of Anli Reservoir in Qixia City, Shandong Province as the study area, fusion images were conducted to combine the PAN and multispectral images of ZY-3 satellite in one image using Gram-Schmidt （GS） spectral sharpening method, hue-saturation-value（HSV） transform, and principal component analysis （PCA） transform. Gully parameters were interpreted from the fusion images, and the accuracy was assessed by comparing the gully parameters interpreted from fusion image against those derived from a ground survey. The results showed that the fusion images constructed using the three methods have not only enriched spectral information but also enhanced spatial texture information. The gully mapping accuracy from the three fused images were much higher than those from the original multispectral image. Compared with the field measurement, the accuracy of gully map interpretated from GS spectral sharpening fusion image was the highest among all the three methods, with correlation coefficients of gully length and gully area between those of field measurement and those of mapping from fusion image being 0.984 and 0.991, respectively. This demonstrated that GS spectral sharpening method was the most suitable image fusion method for gully mapping.

关 键 词：冲沟参数 Gram-Schimdt HSV变换 PCA变换 图像融合 资源三号卫星 

分 类 号：S157[农业科学—土壤学]