CBERS-02 IRMSS 热红外数据及UTAE 算法在城市热岛研究中的应用

Abstract

This paper discusses the use of the Algorithm of Urban Thermal Anomaly Extraction (UTAE) in extracting information on urban heat islands (UHI) from CBERS-02 Infrared Multispectral Scanner (IRMSS) data. The algorithm uses a robust threshold estimation method that scans the entire image through a moving window (3×3 to 9×9 pixels) and calculates a threshold of mean thermal value +1×standard deviation. It distinguishes between UHI and non-UHI areas and assesses UHI intensity according to the cumulative number of pixels identified as thermal anomalies. Threshold estimation is dynamic and the algorithm can be considered unbiased. This method is used to estimate the area of UHI in Changsha which varies from 27.21 km² to 33.98 km² on 13 February 2004. As to UHI intensity analyses at macro-scale, larger windows should be used to reflect the relationship between urban and rural areas embodied in general development trends, while studies at micro-scale smaller windows should be used to analyze the specific distribution of UHIs and identify relationships among different types of underlying surface, and their driving factors. Finally, we discuss the applicability of the method to data generated by CBERS and its successors.