Edge identification method, system and equipment for agricultural abnormal data and medium

The invention discloses an agricultural abnormal data edge identification method, system and device and a medium, and the method comprises the steps: collecting crop sensing parameter data through a sensing node, and carrying out the abnormal data initial identification of the sensing parameter data through a threshold method, and obtaining sensing data; the sensing data is uploaded to a sink node and temporarily stored in an edge server; the edge server randomly selects a certain type of perception data, and constructs a multivariate data exception index function according to the correlation between data types; and the edge server screens out sensing nodes adjacent to the sensing nodes corresponding to the sensing data of the type, selects two types of sensing data most related to the sensing data of the type from the adjacent sensing nodes, and calculates a comprehensive anomaly index to perform local multivariate data anomaly detection on the sensing data of the type. Aiming at the problem of abnormal data