E-nose Gas Data Classification System and Method Using Neural Networks

摘要

The present invention provides a gas data input unit for receiving electronic nose gas data reconstructed from a denoising auto-encoder, and at least one channel-side vector information of the received electronic nose gas data to a one-dimensional convolution filter set. a gas data classification unit that extracts and encrypts the data based on the data, merges the encrypted at least one or more channel-side vector information, and classifies the received electronic nose gas data through a fully connected layer based thereon This has the effect of increasing the reliability of gas type classification.