Intelligent seal movement detection method based on artificial intelligence

Abstract

The invention relates to an intelligent seal movement detection method based on artificial intelligence, and the method comprises the steps: collecting real-time data of the movement acceleration of a seal in an automatic sealing process, inputting a trained TCN model based on an attention mechanism, obtaining a classification result, judging whether a manual seal movement behavior occurs in the automatic sealing process or not according to the classification result, If so, giving an alarm, otherwise, ending; wherein the training process of the TCN model specifically comprises the steps of acquiring moving acceleration sample data of a seal in the automatic sealing process, creating a sample set, and preprocessing the set; and training the TCN model by using the sample set. Compared with the prior art, the invention has the advantages of automation, high accuracy, good robustness and the like.

Description

technical field [0001] The invention relates to a movement detection technology in the field of intelligent seals, in particular to an artificial intelligence-based movement detection method for intelligent seals. Background technique [0002] The seal is used as the seal for the company to handle internal and external affairs, and it is the certificate for the validity of the company's contracts, official documents, letters of introduction, certification letters and other documents with the nature of legal documents. In the process of using the seal, there are often risks such as unauthorized stamping and illegal stealing, which may bring unnecessary legal disputes and unbearable responsibilities to the seal holder. Therefore, the importance of seal control is self-evident. [0003] The invention of the existing smart seal technology proposes a smart seal management solution, which controls the entire process of stamping, realizes no contact in the entire stamping process ...

AI Technical Summary

Problems solved by technology

However, when the automatic stamping is started, the user moves and replaces the stamped document privately. The traditional mobile monitoring method sets the single-axis acceleration threshold of the accelerometer, and judges the movement behavior according to whether the threshold is exceeded. This method first ignores the stamping process. The diversity of mobile behaviors in China also relies on the subjective experience of forecasters, resulting in low robustness and unstable forecasting results

Images