With the growing frequency and diversity of threats targeting critical infrastructure, rapid and reliable intrusion detection has become essential. While video surveillance systems such as CCTV are widely used, radar-based solutions offer key advantages, including long-range detection and robustness to poor visibility conditions. Among these, micro-Doppler (mD) signature classification has gained significant attention, primarily focusing on binary human detection using either feature-based or deep learning approaches. However, such methods often lack generalizability and adaptability to complex environments or diverse object types. This study introduces a hybrid knowledge-based framework for mD signature interpretation. The approach commences with a morphological decomposition of mD signals in outdoor environments, thereby enabling the identification of sub-signal components.