基于不变目标法的DMSP/OLS夜间灯光影像校正

Night Light Image Correction of DMSP/OLS Based on Invariant Target Method

通过社会经济状况的分析,对传统不变目标区域进行了重新选择.以黑龙江牡丹江市作为不变目标区域,以夜间灯光数据及统计数据为基础,选择F12_1996、F14-15_2003和F1_2010共3期辐射定标数据对中国区域的34期夜间灯光影像进行校正.提取中国区域31省市夜间灯光强度,通过与统计GDP数据建立指数、线性、对数、二次项和幂函数回归模型,并计算统计GDP与预测GDP之间的相关性与平均相对误差,选出符合各省的最佳预测模型.结果表明:①通过校正解决了长时间序列夜间灯光数据的不连续、不稳定问题;②校正后的31个行政区与统计GDP构建的预测模型优于校正前预测模型.模型之间的相关性均在0.95以上,平均相对误差在20%,其中部分地区低于1%,如安徽和福建等地;③经过校正后的灯光数据,除西北地区的部分地级市GDP被严重高估,其他地级市的误差在30%以内.

Based on the analysis of social and economic situation,this paper makes a new choice to the traditional invariable target area taking Mudanjiang city of Heilongjiang province as the invariant target area,the night light data and statistical data as the basis.Three phases of radiometric calibration data(F12_1996,F14-15_2003and F16_2010)were selected to correct the 34phases of night light images in China.The night light intensity of 31provinces and cities in China was extracted,and regression models of exponential,linear,logarithmic,quadratic and power function were established by combining with statistical GDP data.The correlation and average relative error between statistical GDP and predicted GDP were calculated,so as to select the best prediction model that conforms to the provinces.The results show that:①The problem of discontinuity and instability of night-time light data of long time series is solved by correction;②The prediction model constructed by the 31administrative regions after correction and GDP statistics is superior to the model of all administrative regions before correction.The correlation between the models is above 0.95,the average relative error is 20%,and some regions are less than 1%,such as Anhui and Fujian;③After the correction of the lighting data,except for some prefecture level cities in Northwest China whose GDP is seriously overestimated,the error of other prefecture level cities is less than 30%.