The application relates to the technical field of
water quality detection, and discloses a
karst water quality detection method and
system based on underground water
dynamic data. The method comprises the following steps: acquiring a multi-dimensional sensing sequence composed of
water level,
turbidity, pH and
calcium concentration; aligning the
water level and the
turbidity and performing difference, to obtain a change rate and a change sequence; calculating cross-correlation to extract a time
delay feature, and modeling to obtain a
coupling strength index; establishing a
carbonate balance model based on the pH and the
calcium concentration, to obtain a constraint parameter, combining the
coupling strength to perform state
estimation, and obtaining an adjusted
coupling vector; performing high-frequency focusing on the multi-dimensional sequence, to obtain a high-frequency component, evaluating a driving strength, and obtaining a regulation degree value; fusing
water level to reconstruct a
dynamic feature, and performing deep inversion to obtain a deterioration trend; performing energy weighted fusion to calculate statistical moments, to obtain an inversion parameter set; performing drift
verification to obtain a
feature vector, and finally performing space-time fusion to output a dynamic detection result. The method can improve the accuracy of
karst underground
water quality detection.