Hydraulic engineering management data processing method and system based on cloud computing

Abstract

The invention discloses a water conservancy project management data processing method and system based on cloud computing, and relates to the technical field of data processing. Real-time video frames are input into a feature extraction network, a saliency thermodynamic diagram is generated, a region of interest is cut out, the real-time video frames are input into a feature detection network, and water level line pixel coordinates are output and mapped into structured water level values; the method comprises the following steps: performing time sequence data stream shunting to obtain non-stationary fluctuation data and time sequence void data, shunting the non-stationary fluctuation data to a first processing channel to obtain an environmental noise entropy vector, shunting the time sequence void data to a second processing channel, obtaining a final complementation value in combination with the environmental noise entropy vector and historical hydrological data, and generating a reconstructed time sequence data stream. And marking statistical outliers for the reconstructed time sequence data flow, logically matching the statistical outliers with a historical association rule set, calculating a physical deviation index, and marking the statistical outliers as equipment fault data or effective mutation data in a classified manner based on a logical matching result and the physical deviation index.

Images