基于Landsat-8 OLI/TIRS数据的地表温度反演算法对比研究

Comparison of land surface temperature retrieval algorithm based on the Landsat8 OLI/TIRS data

关键词:  地表温度  Landsat-8  覃志豪单窗算法  劈窗算法  呼和浩特

Key words:  land surface temperature  Landsat-8  mono-window algorithm by Qin Zhihao  split-window algorithm  Hohhot

摘要:

【目的】 评估不同反演算法在卫星遥感地表温度数据获取中的性能差异，特别是针对呼和浩特地区在特定时间的应用。【方法】 文章基于Landsat-8 OLI/TIRS数据，选用3种不同的反演算法（覃志豪单窗算法、Offer Ronzenstein劈窗算法和Jiménez-Mu?oz劈窗算法），对呼和浩特区域的地表温度进行反演；随后对反演结果进行精度检验和敏感性分析，以评估不同算法在相同条件下的表现。【结果】 经过对比分析，覃志豪单窗算法在呼和浩特区域的地表温度反演中展现了最高的精度，其平均相对误差为3.68%；相比之下，Offer Ronzenstein劈窗算法和Jiménez-Mu?oz 劈窗算法的平均相对误差分别为4.22%和6.60%。在敏感性分析方面，Jiménez-Mu?oz 劈窗算法对水汽含量的变化最不敏感，其次是Offer Ronzenstein劈窗算法，而覃志豪单窗算法则表现出较高的敏感性，并仅适用于特定范围内的水汽含量。【结论】 在呼和浩特地区的地表温度反演中，覃志豪单窗算法因其较高的精度而成为优选算法。然而，不同算法在敏感性和适用范围上的差异表明，在实际应用中，需要根据具体的数据特征和区域条件选择合适的反演算法。此外，进一步的研究可以探索更多算法以提高地表温度反演的准确性和普适性。

Abstract:

Purpose This study aims to evaluate the performance differences of different inversion algorithms in satellite remote sensing land surface temperature（LST）data acquisition，especially for the Hohhot area at a specific time.Method Based on Landsat-8 OLI/TIRS data，three different inversion algorithms（mono-window algorithm by Qin Zhihao，Offer Ronzenstein split-window algorithm，and Jiménez-Mu?oz split-window algorithm）were selected to invert the surface temperature in the Hohhot region. Subsequently，accuracy tests and sensitivity analysis were conducted on the inversion results to evaluate the performance of different algorithms under the same conditions.Result After comparative analysis，the mono-window algorithm by Qin Zhihao demonstrated the highest accuracy in surface temperature inversion in the Hohhot region，with an average relative error of 3.68%. In contrast，the average relative errors of the Offer Ronzenstein split-window algorithm and the Jiménez-Mu?oz split-window algorithm were 4.22% and 6.60%，respectively. In terms of sensitivity analysis，the Jiménez-Mu?oz split-window algorithm was the least sensitive to changes in water vapor content，followed by the Offer Ronzenstein split-window algorithm，while the mono-window algorithm by Qin Zhihao showed higher sensitivity and was only applicable to a specific range of water vapor content.Conclusion The mono-window algorithm by Qin Zhihao is the preferred algorithm for surface temperature inversion in the Hohhot area due to its high accuracy. However，the differences in sensitivity and applicability of different algorithms indicate that in practical applications，it is necessary to select appropriate inversion algorithms according to specific data characteristics and regional conditions. In addition，further research can explore more algorithms to improve the accuracy and generalizability of surface temperature inversion.