Article identification method, device and equipment applied to self-service vending device

The invention provides an article identification method, device and equipment applied to a self-service vending device. The method comprises the following steps: acquiring a visual identification result, an article gravity change value of the visual identification result and an overall actual gravity change value of the unmanned vending device; in response to the fact that the article gravity change value of the visual recognition result is not matched with the overall actual gravity change value, generating n prediction results of reduced articles in the unmanned vending device; calculating confidence coefficients respectively corresponding to the n prediction results; selecting m kinds of prediction results with confidence coefficients meeting a first condition from the n kinds of prediction results; obtaining article gravity change values respectively corresponding to the m prediction results; and determining a target prediction result that the article gravity change value in the mprediction results meets