A Multi-Camera Image Processing and Visualization System for Train Safety Assessment

摘要

In this paper we present a machine vision system to efficiently monitor,analyze and present visual data acquired with a railway overhead gantryequipped with multiple cameras. This solution aims to improve the safety ofdaily life railway transportation in a two- fold manner: (1) by providingautomatic algorithms that can process large imagery of trains (2) by helpingtrain operators to keep attention on any possible malfunction. The system isdesigned with the latest cutting edge, high-rate visible and thermal camerasthat ob- serve a train passing under an railway overhead gantry. The machinevision system is composed of three principal modules: (1) an automatic wagonidentification system, recognizing the wagon ID according to the UICclassification of railway coaches; (2) a temperature monitoring system; (3) asystem for the detection, localization and visualization of the pantograph ofthe train. These three machine vision modules process batch trains sequencesand their resulting analysis are presented to an operator using a multitouchuser interface. We detail all technical aspects of our multi-camera portal: thehardware requirements, the software developed to deal with the high-frame ratecameras and ensure reliable acquisition, the algorithms proposed to solve eachcomputer vision task, and the multitouch interaction and visualizationinterface. We evaluate each component of our system on a dataset recorded in anad-hoc railway test-bed, showing the potential of our proposed portal for trainsafety assessment.