DMSP/OLS和VIIRS/DNB夜间灯光影像的校正及拟合

运用不变目标法对DMSP/OLS影像数据进行了相互校正、连续性校正和过饱和校正，以校正后的DMSP/OLS数据为参考对VⅡRS/DNB数据进行了重分类，利用两者影像数据在时间与空间中的重叠区域拟合VⅡRS/DNB数据，并对结果进行了评价，目的是获得长时间序列连续稳定的夜间灯光数据。同时建立了新疆区域校正后影像数据与社会经济指标中年末人口总数、生产总值、平均每天耗电量、建设用地之间的关系，探讨了夜间灯光数据在长时间序列下模拟社会经济参量的能力。结果表明，运用不变目标法校正后的夜间灯光影像更稳定、更连续，缩小了VⅡRS/DNB数据与DMSP/OLS数据像元DN值量级差异，使长时间序列下的灯光数据具有可比性。校正后长时间序列夜间灯光数据与社会经济参量的相关性高，说明校正后的长时间序列灯光影像数据在省级尺度上与社会经济指标契合。

Correction and fitting of night light images of DMSP/OLS and VIIRS/DNB

In this paper, the DMSP/OLS image data is corrected by mutual invariance, continuity correction and oversaturation correction. The corrected DMSP/OLS data is used as a reference to reclassify the VⅡRS/DNB data. The two image data are used in time. The VⅡRS/DNB data fitted to the overlapping area in the space and the results are evaluated to obtain long-term continuous and stable nighttime lighting data. At the same time, the relationship between image data and social and economic indicators of Xinjiang regional corrections, the total population at the end of the year, total production value, average daily power consumption, construction land use are established, and the ability of nighttime lighting data to simulate social and economic parameters under long-term sequence is explored. The results show that the nighttime light image after correction by the invariant object method is more stable and continuous, which reduces the magnitude difference of the DN value between the VⅡRS/DNB data and the DMSP/OLS data pixel, making the light data under long-term sequence comparable. The correlation between long-term sequential nighttime lighting data and socioeconomic parameters after correction is high, indicating that the corrected long-term sequence light image data is consistent with socioeconomic indicators at the provincial level.