铁路客运场景下基于图像搜索的遗失物品查找方法

Image Search-Based Method for Finding Lost Items in Railway Passenger Transport Scenarios

当前铁路客运场景下的遗失物品查找方法效率低下,影响了旅客的出行体验,给各车站的生产经营造成了困扰。为创新铁路客运场景下的遗失物品查找方式,在分析铁路客运遗失物品查找需求与难点的基础上,结合人脸识别以及深度学习的前沿技术成果,建立了一种基于图像搜索的遗失物品查找框架,设计了面向铁路客运场景的安检遗失物品查找方案以及非安检遗失物品查找方案。研究结果表明,该方法可进一步提高铁路客运运营的智能化水平,优化遗失物品的查找效率,在跨模态检索测试中有较高的检索精度,但在部分类别中的检索结果存在误差。基于研究结果,从算法改进与模型微调策略等方面进行了展望。

At present,the method for finding lost items in the railway passenger transport scenario is inefficient,which affects the travel experience of passengers and causes trouble to the production and operation of stations.In order to innovate the method for finding lost items in the railway passenger transport scenario,the needs and difficulties of finding lost items in the railway passenger transport scenario were analyzed.Combined with the cutting-edge technology achievements of face recognition and deep learning,a framework for finding lost items based on image search was established,and the scheme for finding lost items during security check and non-security check periods in the railway passenger transport scenario was designed.The results show that this method can further improve the intelligence level of railway passenger transport operations and promote the efficiency of finding lost items.It has high retrieval accuracy in cross-modal retrieval tests,but there are a few errors in the retrieval results of some categories.Based on the research results,the algorithm improvement and model fine-tuning strategies were forecasted.