论文部分内容阅读
通过遥感获取地表温度(地温,LST)可以弥补气象站LST数据局地性的不足。但受某些因素影响,遥感LST影像存在噪音像元,影响了LST数据的应用。本文提出了一种基于高程-温度回归关系的空间重建算法,对2008年青藏高原MODIS LST影像异常低值像元进行了重建,得到空间完整的LST时间序列。分别将原始LST和重建LST与研究区62个气象站最高气温进行对比分析,结果表明,重建前后LST与最高气温一致性很好,相关系数分别为0.88和0.91,各气象站平均相关系数达到0.898,LST一般高于气温,平均偏差为6.98℃。地气温差与海拔、土地覆盖、季节和噪音水平有关。海拔越高,地气温差越大;植被覆盖越好,地气温差越小;温度越高,温差越大。研究还发现,重建后的数据在森林覆盖区仍然残留噪音,这需要结合LST的质量控制信息做进一步研究。
The remote sensing of surface temperature (ground temperature, LST) can make up for lack of local station LST data. However, influenced by some factors, there are noise pixels in remote sensing LST images, which affects the application of LST data. In this paper, a spatial reconstruction algorithm based on elevation-temperature regression is proposed to reconstruct the low-value pixels of MODIS LST images in Qinghai-Tibet Plateau in 2008, and a complete LST time series is obtained. The comparison between the original LST and the reconstructed LST and the maximum temperature of 62 meteorological stations in the study area shows that the consistency between the LST and the maximum temperature before and after reconstruction is very good with the correlation coefficients of 0.88 and 0.91 respectively and the average correlation coefficients of all weather stations reach 0.898 , LST is generally higher than the temperature, the average deviation of 6.98 ℃. The difference in ground temperature is related to elevation, land cover, season and noise level. The higher the altitude, the greater the temperature difference; the better the vegetation coverage, the smaller the temperature difference; the higher the temperature, the greater the temperature difference. The study also found that the reconstructed data still has residual noise in the forest cover area, which needs to be further studied in combination with the LST quality control information.