Research on CRNN based recognition method for cigarette uplink code
In the cigarette logistics system,the cigarette traceability label contains part of the key sorting information not included in the one-dimensional barcode,which can greatly improve the sorting efficiency.The crucial aspect is high-precision label recognition.To accurately recognize cigarette labels,we propose an improved CRNN-based traceability label recognition network,termed RA-CRNN.This method's feature extraction,inspired by ResNet,introduces a residual structure and enhances recognition accuracy via attention and gating mechanisms.Comparative experimental analyses demonstrate that this improved algorithm surpasses other advanced text recognition algorithms in accuracy.For instance,in a Ningbo cigarette factory,the traceability label's identification accuracy reached 97.87％,fulfilling industrial automation requirements for cigarette traceability label identification.
万方数据
维普
