中文摘要：

以深圳市区一景QuickBird Pan波段遥感图像（总行列数为26 574×28 606,对应15.944 km×17.164km的实地范围）为研究对象,利用深圳市区GPS虚拟参考网络定位系统,以厘米级精度测定了65个在图像上易于识别的地面控制点三维坐标,通过65个控制点之间真实距离和在变形图像上对应距离之间的偏差,在分析线性阵列传感器成像原理,遥感图像像点位移影响因素及规律的基础上,利用平均残差平方和准则,系统分析了Quick-Bird Pan波段遥感图像的变形规律和定量估测模型。研究成果对高空间分辨率遥感图像几何精纠正算法研究具有应用价值。

英文摘要：

Three-dimensional coordinates of 65 ground control points（GCPs） easily recognized on images were measured using one scope of QuickBird remote sensing image of panchromatic band（total rows and columns are 26 574 and 28 606,the corresponding real range of 15.994 km×17.164 km） in Shenzhen urban area as study object by means of the virtual GPS reference system network.According to the deviation between the real distance of 65 GCPs and their corresponding distance on the deformed image,the deformation laws and the quantitative description model of QuickBird remote sensing images of panchromatic band were analyzed systematically by the rules of mean residual sum of squares based on the analysis of the imaging principle of linear array sensor and the factors influencing the pixel displacement of remote sensing images.The method could be applied to the study of geometric calibration model for high-resolution remote sensing images.