Low cost baggage tracking, through machine learning and image processing.

开发一种创新的行李标签识别系统，该系统利用低成本相机的改进功能，以及先进的图像分析和机器学习技术，在从办理登机手续到最终目的地交付给乘客的整个过程中，对行李进行单独识别。行李处理不当每年给航空运输业造成23亿美元的损失，而欺诈和夸大索赔正成为一个日益严重的问题。为了减少处理不当和欺诈行为，代表265家航空公司的行业协会国际航空运输协会（IATA）通过了一项决议（753），要求到2018年6月，其成员在行李处理过程（登机、装载、转移和交付）的关键交换点跟踪托运行李。然而，到目前为止，世界上很少有航空公司和机场拥有必要的基础设施来履行这些跟踪义务，主要是由于成本、缺乏准确性和典型行李大厅的物理限制。围绕紧凑、低成本的相机、先进的图像分析和机器学习构建的行李标签识别系统应该能够解决这些传统限制，并为航空公司和机场提供可靠且负担得起的选择，以进行更广泛的行李跟踪。

The development of an innovative bag tag recognition system that uses the improved capabilities of low cost cameras and advances in image analysis and machine learning to individually identify bags as they are handled on their journey from check-in through to delivery to the passenger at their final destination. Mishandled baggage cost the air transport industry $2.3billion per annum, while fraudulent and exaggerated claims are becoming a growing problem. In an attempt to reduce mishandling and fraud, IATA, the trade association representing 265 airlines, has passed a resolution (753) that requires, by June 2018, that their members track checked-in baggage at key exchange points in the baggage handling process (check-in, loading, transfer and delivery). However, as of today, very few airlines and airports worldwide have the necessary infrastructure to meet these tracking obligations, primarily due to cost, lack of accuracy and the physical constraints of a typical baggage hall. A bag tag recognition system, built around compact, low cost cameras, advanced image analysis and machine learning should be able to address these legacy restrictions and provide airlines and airports with a credible and affordable option for more extensive baggage tracking.