<p>This study proposes a concrete piezoelectric sensing detection method based on mathematical morphology and fractal theory, which effectively monitors and quantitatively assesses the dynamic evolution process of damage cracks in concrete structures under impact loads. The main contents of this study are as follows: First, it establishes the mapping relationship between the dynamic evolution of damage cracks in concrete under impact loads and the characteristics of piezoelectric time-domain signals for the first time. Through systematic research on the evolution law of peak characteristic parameters of signals in each stage of crack propagation, the intrinsic correlation between the degree of damage and acoustic signals is revealed. Second, it systematically conducts morphological parameter analysis of piezoelectric sensing signals and calculates the morphological fractal dimension (MFD) of piezoelectric signals. Third, it innovatively constructs an intelligent structural damage recognition model integrating morphological fractal theory and artificial neural network (ANN), and conducts a systematic comparative analysis with the traditional wavelet packet transform (WPT) method, verifying the effectiveness of the proposed MFD-ANN intelligent recognition model in this paper. The research results show that the signal corrosion algorithm based on mathematical morphology can significantly enhance the contrast of the steepness characteristics of wave peaks at different damage stages, thereby more effectively capturing the self-similarity characteristics of signal waveforms. Compared with the traditional wavelet packet transform method, the intelligent recognition model established by integrating fractal features and neural networks has a higher recognition accuracy rate for the degree of damage.</p>