With our study presented at the SPIE conference “Earth Resources and Environmental Remote Sensing/GIS Applications” last year, we discussed a concept for change analysis in sequences of SAR sub-aperture images. The main aspect of this concept is to investigate an adaption of our approach for incoherent time series change analysis on such short-term time series data. In a first step, sub-aperture amplitude images of two maritime scenes were calculated leading to time series stacks being considered as input for our incoherent change detection method. As output, so-called ActivityMaps (AMs) were constructed aiming on a recognition of high activity areas. Focusing on short-term time series, such areas are caused by Ground Moving Targets (GMTs) which denote objects that were in motion during the image acquisition. With respect to the maritime scenes considered in this study, GMTs might be ships, cars, trucks, cranes with moving components, etc. It was observed, that GMTs show different signatures in the AMs, depending for example on their size and their velocity. In the paper at hand, we link to this previous study by discussing different features being reliable for a later categorization of the detected change objects. Moreover, it is investigated, whether High Activity Objects (HAOs) are of solely interest, or, if other change objects have to be included. The relevance of the discussed features to produce categories being clearly distinguishable from each other is tested by an unsupervised clustering procedure. As test data, a TerraSAR-X (TSX) Staring Spotlight (ST) Single Look Complex (SLC) image of Rotterdam (NED) was used.