Industrial anomaly detection method by fusion of self-coding feature extraction and continuous immunization mechanism

Timely and accurate detection of anomalies during the industrial production process is crucial for ensuring production safety and efficiency. This paper proposes an industrial anomaly detection method that integrates autoencoder feature extraction and a continuous immune mechanism. The autoencoder is employed to extract features from industrial data, excavating the latent features and patterns within the data, reducing the data dimension, and the Adam optimizer is adopted to realize the extraction of feature information. A continuous immune mechanism is constructed through the Dendritic Cell Algorithm (DCA) to achieve real-time monitoring and dynamic response to industrial anomalies. The experimental results demonstrate that this method has a high accuracy rate and a low false positive rate in industrial anomaly detection, and it can effectively adapt to the dynamic changes of industrial data, providing a reliable anomaly detection solution for industrial production.