基于零样本学习的集装箱箱号识别系统研究

Research on Container Number Recognition System Based on Zero-shot Learning

针对传统集装箱号识别方法对于新出现的没有标签的集装箱无法进行准确识别问题,介绍一种用于集装箱号自动识别的基于语义自编码器的零样本学习方法。该方法可实现对陌生集装箱号的准确识别;给出一种基于该方法的箱号识别系统及其智能视觉识别系统、后端集中管控系统、web前端系统组成,可为集装箱号识别解决方案提供参考。

Aiming at the problem that the traditional container number identification method cannot accurately identify the newly emerging containers without labels,a zero-shot learning method based on Semantic Auto-Encoder for automatic identification of container number is introduced.This method can realize the accurate identification of unfamiliar container number;and a container number identification system based on this method,its intelligent visual recognition system,back-end centralized management and control system,and web front-end system are presented,which can provide a reference for container number identification solutions.