This invention provides a reservoir dam
safety monitoring system, the core mechanism of which includes: using a pole-symmetric mode
decomposition algorithm to perform multi-scale
decomposition on the dam structure
response time series data to obtain several intrinsic mode functions; adaptively filtering each intrinsic mode function through
permutation entropy; performing multi-scale grouping and weighted reconstruction of effective structural feature
modes through dispersion entropy to obtain multiple sets of reconstructed sequences at different scales; constructing a multi-scale entropy
feature vector corresponding to the dam structure response;
coupling and analyzing the entropy feature sequence obtained by the same
processing of the running command
time series data with the multi-scale entropy
feature vector to output the dam safety state coefficient; performing target detection, tracking, and
behavioral risk analysis on the panoramic video
stream data of the dam area to output the personnel
behavioral risk coefficient; and performing a fusion
risk assessment based on the dam safety state coefficient and the personnel
behavioral risk coefficient to calculate the comprehensive
risk level and trigger the corresponding level of prevention and control early warning.