Automatic glass bottle detection system and method thereof

The invention relates to the field of intelligent detection, and particularly discloses an automatic glass bottle detection system and a method thereof.The artificial intelligence technology based on a deep neural network model is adopted to obtain a detection image of a finished glass bottle to be detected, and a feature map is obtained through a convolutional neural network; and after the feature map is segmented, deep feature extraction is carried out through a hybrid convolution module and a three-dimensional convolution module, so that a classification result used for representing whether the quality of the finished glass bottle to be detected reaches the standard or not is obtained. Furthermore, the automatic detection process can be realized, and the production efficiency, the accuracy and the quality control level are improved. 本申请涉及智能检测领域，其具体地公开了一种自动化玻璃瓶检测系统及其方法，其采用基于深度神经网络模型的人工智能技术，获取待检测成品玻璃瓶的检测图像，通过卷积神经网络得到特征图，将特征图切分后再通过混合卷积和三维卷积模块进行深度特征提取，以得到用于表示待检测成品玻璃瓶质量是否达标的分类结果。进而，可以实现自动化的检测过程，提高生产效率、准确性和质量控制水