Making Full Use of the Landsat7 SLC-Off ETM plus Data for Urban Thermal Environment Monitoring

Landsat program has provided the well calibrated and relatively high spatial resolution data of the Earth's surface, since the launch of its first satellite (Landsat1) in 1972. The continuity of this program is ensured thanks to the operation of Landsat8 (launched on 11 February 2013). According to the data open policy announced in 2008, all new and archived Landsat data held by the U.S. Geological Survey are freely accessible through the internet to general users worldwide. Unfortunately, the scan-line corrector (SLC) for the Landsat7 ETM+ sensor failed permanently on 31 May 2003. The SLC-off problem limits the quantitative application of ETM+ data, and results in a period of data gap consequently. Therefore, resolving the problem of Landsat7 SLC-off ETM+ imagery is a valuable issue. At present, several approaches have been proposed. However, a few researches have been done to recover the thermal imagery of Landsat7 SLC-off ETM+, while relatively much attention has been paid to the recovering of multispectral bands. Considering the merits of thermal imagery for monitoring urban thermal environment, we evaluated the approaches to recover the thermal band of Landsat7 SLC-off ETM+. Methods in this paper are mainly two current approaches, including the adaptive window linear histogram match (AWLHM) and the spectral and temporal based method (named STM) which is a simple but effective method. Case study shows that with different band or band combination as explanatory variable(s) the modified AWLHM performs well, according to error indicators. In this paper, three error measurements are calculated, which specifically are the universal image quality index (UIQI), the relative mean absolute error (MAE(r)), and the relative root mean square error (RMSEr). Meanwhile, the STM method presents slight advantages over the modified AWLHM methods. Accordingly, most parts of SLC-off affected pixels are able to be recovered properly, confirming that the recovered thermal band can be applied in urban thermal environment issues. We assume that more acquisitions can provide relatively more comprehensive details about urban thermal environment. The importance of the multiple recovered SLC-off thermal imageries is obvious. Accordingly, the significance of the open policy of Landsat series is demonstrated in this paper, in view of monitoring urban thermal environment. Finally, case study indicates that the improvement over the areas with surface heterogeneity is a difficult but worthy task.