Convolutional neural network-based method for image recognition of civil aviation checked luggage

To establish a hierarchical protection system for civil aviation checked luggage and effectively address luggage security challenges, this study proposes an image recognition method based on a Convolutional Neural Network (CNN). First, luggage images were preprocessed using binarization and gradient sharpening techniques to enhance their suitability for recognition. Second, the processed images were segmented by analyzing grayscale characteristics and noise distribution to isolate critical target regions. Finally, key features were extracted from these regions, and the CNN was employed to compute similarity metrics between target images and reference dataset images, thereby determining luggage classification results and enabling hierarchical protection. Experimental results demonstrate that the proposed method achieves a recognition accuracy exceeding 94%.