Dynamics of spatio-temporal objects is desirable for change monitoring. Analyzing and understanding the dynamics of these objects allow discovering geographical processes that can occur. Change detection by using time series of satellite images can help to understand the dynamics of landscape. Spatiotemporal objects in these data are often dynamic by nature. In a given time, state, geometric form, type and identity of object can involved and several relationships with others objects may be exist. In this paper, we highlight the complexity of the dynamics of spatio-temporal object and we propose a conceptual model enabling us to understand the dynamics of geographical processes such as deforestation, desertification, urbanization, etc. through times series of satellites images. Such model is dependent on the use of ontologies in different levels of granularity.