Mining Multiple Satellite Sensor Data Using Collaborative Clustering

In recent years, satellite sensor data have become easier to acquire. Several different satellite systems are now available and produce a large amount of data used for Earth observation. To better grasp the complexity of the Earth surface, it became usual to use different images from different satellites. However, it is generally difficult to predict the potential gain of using multisource satellite sensor data before actually acquiring the data. In this paper, we present a simulation approach to create different views of remote sensing sensor data according to different satellite characteristics. These different views are then used in a collaborative clustering approach to assess the interest of using these multisource data together. Experiments provide some insights on couple of satellite systems able to leverage the complementary of the sources.