土地整理项目中空间数据的精度分析
|
摘要
土地整理定量化研究中空间数据的精度是倍受关注的问题。在利用3S技术对土地整理量化描述的过程中存在一些急需解决的难题：(1)大多地区没有大比例尺的地形图，如何采用高分辨率卫星影像进行校正；(2)原始数据和内插方法是影响DEM精度的主要因素，但空间数据往往没有真值，如何获取真值，分析由原始数据和内插方法对DEM计算土方量的精度的影响；(3)利用地形图矢量化生成DEM模型进行土方量的计算时，如何应用先进的技术进一步提高土地整理中土方量计算的精度。由于误差是空间数据的一个固有属性，它不能完全被消除，只能不断加以改进以提高空间数据的精度，因此，探讨几何校正、不同的插值方法、原始数据及不同来源的误差对数据精度的影响具有重要意义。
本文对以上的问题采取以下的措施进行解决：以大比例尺的(1∶500和1∶2000)地形图校正的高分辨率遥感卫星影像为基准，分析平面精度达亚米级的GPS测量控制点校正的高分辨率遥感影像的校正精度。构建了可模仿凸凹不平地形表面的实体模型，该模型能够表征山脊、山尖、山涧、丘陵等地形特征；根据体积差减法测量实体模型的体积作为数据精度分析的体积“真值”，该值为13260cm~3，作为采用DEM进行体积计算的基准值；利用精度为0.08cm的三维数字化仪采集间距为1cm的离散数据点，对离散数据点用GIS软件进行内插处理生成DEM模型，通过DEM计算出模型的体积，探讨不同内插方法(IDW、Spline、Natural Neighbor和Kriging)和数据格网间距变化对体积计算精度的影响。用高程精度达厘米级RTK到实地测量检测点，对用地形图扫描矢量化生成的DEM进行检测，根据检测点的实测高程和内插高程的差值建立误差DEM，计算误差体积，进一步提高土方量计算的精度。
结果表明：利用平面精度达亚米级的GPS测量控制点坐标对1米分辨率的遥感卫星影像进行几何校正，能够满足其几何校正的精度要求，可以解决缺乏大比例尺地形图地区的高分辨率遥感卫星影像的几何校正问题；通过对模型数据的分析，数据密度是影响体积精度的主要因素，利用检测点高程和特征点可以提高土方量计算的精度，使土方量计算的精度达到90％以上，数据密度越稀疏，精度提高越显著；采用高程精度达厘米级的RTK(Real Time Kinematic实时动态差分)测量检测点的高程，对用地形图扫描矢量化形成的DEM模型进行检测求出误差体积，可以大大提高土方量的计算精度。
The precision of spatial data is an increasing concern over quantification about land consolidation. There are some impending problems on quantified description of land consolidation using "3S" technology: (1) In many regions where the big scale is absent, how to resolve the problem about the geometric correction of the high precision remote sensing image; (2) the original data and the interpolating method are the two main factors that affect the precision of DEM, however, the spatial data has no true value, the problem is how to obtain the true value and to analyze the effect of the original data and interpolating method on the precision of earthwork; (3) when calculating the earthwork through DEM that are constructed by vectoring the relief map, how to improve the accuracy of the earthwork by advanced technology. Since the error is the inherent attribute of the spatial data, it can not be completely eliminate but to improve the data accuracy step by step. Therefore, it is of important significance to discuss the effect of different interpolating method, how to obtain the original data, and the sources of errors on the precision of data.
In this study, firstly, the high resolution remote sensing image, which was geometrically corrected by the big scale relief map (1:500 and 1:2000), was taken as the foundation to analyze the correction precision of the high resolution remote sensing image that was geometrically corrected by the points of GPS. Secondly, two physical models were constructed, which could model rough surface and describe the features of the earth surface, such as chine, neck, valley and hill, etc. According to difference method, the "true" volume of the physical models were measured, which was 13260cm3 and regarded as the analysis foundation of the precision of the volume that was calculate by DEM. The discrete dada of the models on a regular spacing of 1cm were collected by 3SPACE FASTRAK and the DEM was obtained by interpolating the discrete dada through GIS software. Based on the volume of the models calculated by DEM, different interpolating methods (IDW, Spline, Natural Neighbor, and Kriging) and variation of grid spacing of data were compared and their effects on precision of volume calculation were analyzed. Some reference points were measured by the high precision GPS (RTK) to examine the precision of the DEM. The elevation difference between measured and interpolating data was used to construct the DEM and calculate the error of the volume. And then the precision of the earthwork could be improved.
The results showed that: (1) some governing points, which were measured by submeter precision GPS, were utilized to correct the 1 m precision remote sensing image and it could satisfy the desired precision of geometrical correction, therefore, this method could be used as geometrical correction in the region where was short of the big scale relief map; (2) the density of the original data was the main factor to affect precision of the volume, and the precision of the calculated volume of earthwork could be improved to more than 90% when using the test points and feature points; the sparser the data density were, more significant could be improved; On the base that elevation of test points were measured by the submeter precision GPS, the error of the volume, which was produced by vectoring relief map to
obtain DEM, could be calculated, and the calculation precision of the earthwork could be improved markedly.
本文对以上的问题采取以下的措施进行解决：以大比例尺的(1∶500和1∶2000)地形图校正的高分辨率遥感卫星影像为基准，分析平面精度达亚米级的GPS测量控制点校正的高分辨率遥感影像的校正精度。构建了可模仿凸凹不平地形表面的实体模型，该模型能够表征山脊、山尖、山涧、丘陵等地形特征；根据体积差减法测量实体模型的体积作为数据精度分析的体积“真值”，该值为13260cm~3，作为采用DEM进行体积计算的基准值；利用精度为0.08cm的三维数字化仪采集间距为1cm的离散数据点，对离散数据点用GIS软件进行内插处理生成DEM模型，通过DEM计算出模型的体积，探讨不同内插方法(IDW、Spline、Natural Neighbor和Kriging)和数据格网间距变化对体积计算精度的影响。用高程精度达厘米级RTK到实地测量检测点，对用地形图扫描矢量化生成的DEM进行检测，根据检测点的实测高程和内插高程的差值建立误差DEM，计算误差体积，进一步提高土方量计算的精度。
结果表明：利用平面精度达亚米级的GPS测量控制点坐标对1米分辨率的遥感卫星影像进行几何校正，能够满足其几何校正的精度要求，可以解决缺乏大比例尺地形图地区的高分辨率遥感卫星影像的几何校正问题；通过对模型数据的分析，数据密度是影响体积精度的主要因素，利用检测点高程和特征点可以提高土方量计算的精度，使土方量计算的精度达到90％以上，数据密度越稀疏，精度提高越显著；采用高程精度达厘米级的RTK(Real Time Kinematic实时动态差分)测量检测点的高程，对用地形图扫描矢量化形成的DEM模型进行检测求出误差体积，可以大大提高土方量的计算精度。
The precision of spatial data is an increasing concern over quantification about land consolidation. There are some impending problems on quantified description of land consolidation using "3S" technology: (1) In many regions where the big scale is absent, how to resolve the problem about the geometric correction of the high precision remote sensing image; (2) the original data and the interpolating method are the two main factors that affect the precision of DEM, however, the spatial data has no true value, the problem is how to obtain the true value and to analyze the effect of the original data and interpolating method on the precision of earthwork; (3) when calculating the earthwork through DEM that are constructed by vectoring the relief map, how to improve the accuracy of the earthwork by advanced technology. Since the error is the inherent attribute of the spatial data, it can not be completely eliminate but to improve the data accuracy step by step. Therefore, it is of important significance to discuss the effect of different interpolating method, how to obtain the original data, and the sources of errors on the precision of data.
In this study, firstly, the high resolution remote sensing image, which was geometrically corrected by the big scale relief map (1:500 and 1:2000), was taken as the foundation to analyze the correction precision of the high resolution remote sensing image that was geometrically corrected by the points of GPS. Secondly, two physical models were constructed, which could model rough surface and describe the features of the earth surface, such as chine, neck, valley and hill, etc. According to difference method, the "true" volume of the physical models were measured, which was 13260cm3 and regarded as the analysis foundation of the precision of the volume that was calculate by DEM. The discrete dada of the models on a regular spacing of 1cm were collected by 3SPACE FASTRAK and the DEM was obtained by interpolating the discrete dada through GIS software. Based on the volume of the models calculated by DEM, different interpolating methods (IDW, Spline, Natural Neighbor, and Kriging) and variation of grid spacing of data were compared and their effects on precision of volume calculation were analyzed. Some reference points were measured by the high precision GPS (RTK) to examine the precision of the DEM. The elevation difference between measured and interpolating data was used to construct the DEM and calculate the error of the volume. And then the precision of the earthwork could be improved.
The results showed that: (1) some governing points, which were measured by submeter precision GPS, were utilized to correct the 1 m precision remote sensing image and it could satisfy the desired precision of geometrical correction, therefore, this method could be used as geometrical correction in the region where was short of the big scale relief map; (2) the density of the original data was the main factor to affect precision of the volume, and the precision of the calculated volume of earthwork could be improved to more than 90% when using the test points and feature points; the sparser the data density were, more significant could be improved; On the base that elevation of test points were measured by the submeter precision GPS, the error of the volume, which was produced by vectoring relief map to
obtain DEM, could be calculated, and the calculation precision of the earthwork could be improved markedly.
引文
[1] 王军，余莉，罗明，翟刚．土地整理研究综述．地域研究与开发，第22卷第2期2003年4月
[2] 罗明，王军．中国土地整理的区域差异及对策．地理科学进展，2001，20(2)：97-103
[3] 吴次芳，陈美球．乡村土地整理的若干技术问题探讨．中国土地科学，1997(4)：41～45
[4] ErichWei β(贾生华译)．联邦德国的乡村土地整理．北京：中国农业出版社，1999
[5] 徐建春．联邦德国乡村土地整理的特点及启示．中国农村经济，2001(6)：75～80
[6] 鲍海君，吴次芳，叶艳妹，童菊儿，汪峰．土地整理中田块设计和“3S”技术应用研究．农业工程学报，2002年1月第18卷第1期
[7] 李元．国土资源知识读本．北京：中国大地出版社，1998
[8] 王如渊。土地整理与耕地总量平衡。中国土地，1999，(2)：25～26
[9] 马荣华，仆英峡，黄杏元．数字地球时代“3S”集成的发展．“3S”世界，2001年第4期
[10] 李德仁，龚键雅等．GeoStar——中国人为数字地球设计的GIS软件．“3S”世界，2001年第4期
[11] 鲍海君，吴次芳，叶艳妹．“3S”技术支持下的土地整理初探．经济地理，2001(增刊)：21～25。
[12] Ackermann, F.Techniques and strategies for DEM generation, digital photogrammetry: an addendum to the Manual of Photogrammetry, ASPRS, 1996
[13] 李志林，朱庆．数字高程模型．武汉大学出版社。2001年7月
[14] 王万茂．土地整理与可持续发展．国土经济，2002年第12期
[15] 鹿心社．论中国土地整理的总体方略．农业工程学报，2002年1月第18卷第1期
[16] 承继成，郭华东，史文中等．遥感数据的不确定性问题．科学出版社，2003年
[17] 于丽娜．“3S”技术在土地整理项目竣工验收中的应用研究．硕士论文。中国农业大学，2003年
[18] 朱述龙，张占睦．遥感图像获取与分析．科学出版社
[19] 吴连喜．多源遥感数据融合评价的理论与实践．博士论文．中国农业大学，2002，6
[20] 章孝灿，黄智才，赵元洪。遥感数字图像处理。浙江大学出版社。
[21] 王利，郑玮．RTKGPS测量中坐标转换参数求解若干方法讨论．西安工程学院学报，2002，24(1)：63～66
[22] 李德仁，袁修孝．误差处理与可靠性理论．武汉大学出版社
[23] 简灿良．1∶2000数字高程模型(DEM)在高速公路设计中的应用．测绘通报，2001年第2期
[24] Zhilin, Li. A comparative study of the accuracy of digital terrain models (DTEMs) based on
various data Models. ISPRS Journal of Photogrammetry and Rrmote Sensing, 1994, 49(1):2～11
[25] Haining, R. Spatial Data Analysis in the Social and Environ-mental Sciences. Great Britain: Cambridge University Press, 1990, 291～312
[26] 柯正谊，何建邦，池天河．数字地面模型．北京：中国科学技术出版社，1993；57～58
[27] 王斌．DEM土方量计算的误差探讨．铁路航测，1999年第三期
[28] 汤国安，龚健雅，陈正江等．数字高程模型地形描述精度量化模拟研究，测绘学报．2001，30(4)：361～365
[29] 柳长顺，杜丽娟．Arcview在土地整理项目土方量计算中的运用．农业工程学报，第19卷，第二期，2003年3月
[30] Blace, A.E. Determination of Optimum sampling Interval in Grid Digital Elevation Models(DEM) Data Acquisition. Photogrammetric Enigieering and Remote Sensing, 1987, 53:323～330
[31] 刘友光等．工程中数字地面模型的建立于应用及大比例尺数字测图．武汉测绘科技大学出版社，1997
[32] 王茜，薛怀平，吴胜军，徐新刚．利用地形图批量生产DEM数据的方法．上海师范大学学报(自然科学版)。第32卷第2期2003年6月
[33] 唐新明，林宗坚，吴岚．基于等高线和高程点建立DEM的精度评价方法探讨．遥感信息1999年第55卷 第三期
[34] 黄劲风．场地平整中的测量和计算—用PC—E500电子手簿计算土方量．北京测绘，1995，(3)：38～40
[35] 张光辉．快速计算土方量的方法．测绘通报．1997．(5)：23～24
[36] 刘桦．土方量的表格法测算．测绘通报．2000．(4)：64～65
[37] 龚剑文．地图量算．北京：测绘出版社，1991．180-190，300-307
[38] 谭兵，蒋定华，许素芹．一种顾及地形特征的DEM内插方法．测绘学院学报，第18卷增刊，2001年9月
[2] 罗明，王军．中国土地整理的区域差异及对策．地理科学进展，2001，20(2)：97-103
[3] 吴次芳，陈美球．乡村土地整理的若干技术问题探讨．中国土地科学，1997(4)：41～45
[4] ErichWei β(贾生华译)．联邦德国的乡村土地整理．北京：中国农业出版社，1999
[5] 徐建春．联邦德国乡村土地整理的特点及启示．中国农村经济，2001(6)：75～80
[6] 鲍海君，吴次芳，叶艳妹，童菊儿，汪峰．土地整理中田块设计和“3S”技术应用研究．农业工程学报，2002年1月第18卷第1期
[7] 李元．国土资源知识读本．北京：中国大地出版社，1998
[8] 王如渊。土地整理与耕地总量平衡。中国土地，1999，(2)：25～26
[9] 马荣华，仆英峡，黄杏元．数字地球时代“3S”集成的发展．“3S”世界，2001年第4期
[10] 李德仁，龚键雅等．GeoStar——中国人为数字地球设计的GIS软件．“3S”世界，2001年第4期
[11] 鲍海君，吴次芳，叶艳妹．“3S”技术支持下的土地整理初探．经济地理，2001(增刊)：21～25。
[12] Ackermann, F.Techniques and strategies for DEM generation, digital photogrammetry: an addendum to the Manual of Photogrammetry, ASPRS, 1996
[13] 李志林，朱庆．数字高程模型．武汉大学出版社。2001年7月
[14] 王万茂．土地整理与可持续发展．国土经济，2002年第12期
[15] 鹿心社．论中国土地整理的总体方略．农业工程学报，2002年1月第18卷第1期
[16] 承继成，郭华东，史文中等．遥感数据的不确定性问题．科学出版社，2003年
[17] 于丽娜．“3S”技术在土地整理项目竣工验收中的应用研究．硕士论文。中国农业大学，2003年
[18] 朱述龙，张占睦．遥感图像获取与分析．科学出版社
[19] 吴连喜．多源遥感数据融合评价的理论与实践．博士论文．中国农业大学，2002，6
[20] 章孝灿，黄智才，赵元洪。遥感数字图像处理。浙江大学出版社。
[21] 王利，郑玮．RTKGPS测量中坐标转换参数求解若干方法讨论．西安工程学院学报，2002，24(1)：63～66
[22] 李德仁，袁修孝．误差处理与可靠性理论．武汉大学出版社
[23] 简灿良．1∶2000数字高程模型(DEM)在高速公路设计中的应用．测绘通报，2001年第2期
[24] Zhilin, Li. A comparative study of the accuracy of digital terrain models (DTEMs) based on
various data Models. ISPRS Journal of Photogrammetry and Rrmote Sensing, 1994, 49(1):2～11
[25] Haining, R. Spatial Data Analysis in the Social and Environ-mental Sciences. Great Britain: Cambridge University Press, 1990, 291～312
[26] 柯正谊，何建邦，池天河．数字地面模型．北京：中国科学技术出版社，1993；57～58
[27] 王斌．DEM土方量计算的误差探讨．铁路航测，1999年第三期
[28] 汤国安，龚健雅，陈正江等．数字高程模型地形描述精度量化模拟研究，测绘学报．2001，30(4)：361～365
[29] 柳长顺，杜丽娟．Arcview在土地整理项目土方量计算中的运用．农业工程学报，第19卷，第二期，2003年3月
[30] Blace, A.E. Determination of Optimum sampling Interval in Grid Digital Elevation Models(DEM) Data Acquisition. Photogrammetric Enigieering and Remote Sensing, 1987, 53:323～330
[31] 刘友光等．工程中数字地面模型的建立于应用及大比例尺数字测图．武汉测绘科技大学出版社，1997
[32] 王茜，薛怀平，吴胜军，徐新刚．利用地形图批量生产DEM数据的方法．上海师范大学学报(自然科学版)。第32卷第2期2003年6月
[33] 唐新明，林宗坚，吴岚．基于等高线和高程点建立DEM的精度评价方法探讨．遥感信息1999年第55卷 第三期
[34] 黄劲风．场地平整中的测量和计算—用PC—E500电子手簿计算土方量．北京测绘，1995，(3)：38～40
[35] 张光辉．快速计算土方量的方法．测绘通报．1997．(5)：23～24
[36] 刘桦．土方量的表格法测算．测绘通报．2000．(4)：64～65
[37] 龚剑文．地图量算．北京：测绘出版社，1991．180-190，300-307
[38] 谭兵，蒋定华，许素芹．一种顾及地形特征的DEM内插方法．测绘学院学报，第18卷增刊，2001年9月