An image layer difference index method to extract light area from NPP/VIIRS nighttime light monthly data

Abstract

ABSTRACTNighttime light data have been proven to perform well for economic measurements and built-up area extractions. However, the new generation of monthly NPP/VIIRS data series products, which collected from the Visible Infrared Imaging Radiometer Suite (VIIRS) sensor on the Suomi National Polar-orbiting Partnership (NPP) Satellite contains background values, which relatively restrict the potential application of nighttime images. Therefore, the image layer difference index (LDI) was constructed to obtain high-intensity pixels that have radiance mainly caused by urban light glow. In addition, an empirical threshold was used to extract low-intensity pixels. Thirty-eight counties in Chongqing Province in China were used as an example for this experimental research. To verify the accuracy of the proposed method, a linear regression model, which was based on the quantitative statistics of the pixels, exhibited the coefficient of determination R2 values of 0.980 and 0.947. Considering the interrelation between the light intensity and economic level, statistical data of the county-level economy was also used for further confirmation; the R2 values of the linear regression between the computed indexes were 0.924 and 0.909.