Water quality detection error correction method based on embedded computing

Abstract

The application discloses a water quality detection error correction method based on embedded computing, which comprises the following steps: S1, acquiring multi-source data and pre-processing to form a short window detection queue; S2, initializing an interactive multi-model Kalman filter set at an embedded end and configuring a model transition probability matrix; S3, predicting a detection value change trend and updating an error estimation state to form an error estimation result and a model residual result; S4, performing matching calculation on a residual amplitude and a residual dispersion degree and recursively propagating a model probability weight to a current sampling period to adjust a current probability weight; S5, performing posterior re-estimation on the error estimation result and executing residual gating, performing bias compensation and disturbance compensation on original detection values to form a water quality detection result; and S6, collecting feedback data and incrementally updating error compensation parameters. The application can realize localized identification of water quality detection errors and improve the stability of water quality detection results.

Images