Primary Target Detection Method Based on Connected Domain Filtering in Millimeter-Wave Radar

As society’s demand for privacy and security rises, the demand for human monitoring equipment grows accordingly, especially with the development of physiological feature detection technology based on millimeter-wave radar. In this context, the study of human feature detection based on radar echo signals becomes particularly critical. This paper proposes a detection method for human torso targets. Specifically, we develop a primary target detection method based on connected domain filtering (PCF) on processed echo maps. This method effectively reduces noise while retaining key torso information, demonstrating higher adaptability in target detection for human pose classification and providing an effective means for accurate pose classification.