Optimizing Cloud Detection Index for MODIS Imagery Using Genetic Algorithm

Due to impact of the clouds on precipitation and weather forecasting, cloud detection is one of the important applications of remote sensing. Most of the work on cloud detection is based on empirical experiments. For example, these works are based on low temperature of the cloud, where thermal infrared band is used for the detection. In this paper, Principal Component Analysis (PCA) is used to reduce the dimensionality of the dataset and to identify the most important bands amongst MODIS´s 36 multispectral bands. Using Genetic Algorithm (GA), this work aims to find a more accurate daytime cloud detection index capable to adapt to different satellites.