Hydrology telemetering data anomaly detection method based on spatio-temporal characteristics

The invention discloses a hydrology telemetering data anomaly detection method based on spatio-temporal characteristics. The method includes the following steps of 1) carrying out the sub-sequence division on the hydrology telemetering data by using a sliding window method, and carrying out feature extraction on each group of segmented sub-sequences, and forming a set of two-dimensional points capable of being subjected to static method clustering through combination by a certain algorithm, and performing the dimensionality reduction on the two-dimensional points to make the points be the static data which can be clustered; 2) expanding the K-means method into a three-dimensional space for clustering the hydrology telemetering data; 3) calculating the incidence relation between all hydrology telemetering data clusters to serve as a basis for judging whether the hydrology telemetering data is abnormal or not; and 4) determining whether the hydrology telemetering data sub-sequence of a certain measuring station