Monitoring the biophysical composition changes of urban environments using normalized spectral mixture analysis

Analyzing and monitoring biophysical composition changes of urban environments is an important issue due to the rapid urbanization in China. Unfortunately, the estimation of urban composition using moderate spatial resolution satellite images presents great challenges due to the diversity of urban landscapes. In this paper, we performed NSMA proposed by Wu (2004) to analyze biophysical composition changes of Nanjing city, East China of the two dates using TM and ETM+ images. Three percentage composition images corresponding to the three endmembers (vegetation, impervious surface and water or shade) were acquired by applying a fully constrained SMA to the normalized images. The fraction changes of three kinds of composition were systematically analyzed through associating them with urban growth of Nanjing city. The accuracy of composition images was validated using IKONOS images. Results show that NSMA is able to monitor urban biophysical composition changes effectively and has a considerable accuracy.