VISION BASED TRAIN-WAGON RECOGNITION

摘要

A novel train-wagon recognition system based on copmuter-aided vision will be introduced in this paper. This system finds its application at Hamburg Harbor. It uses a sensor triggered camera and an OCR-System (Optical Character Recognition) for identifying wagon numbers and verifies them to a given list. The application uses a novel algorithm for pre-processing the color images before the data is processed by the OCR. The numbers that are extracted are checked for plausibility by two rails sensors, a weighting machine and an axis counter.