Analysis of models for image processing when implementing an object detection system on a railway track

With the development of railway industry, informatization society and automation many technological processes, it becomes possible to create an automatic control system, diagnostics safety locomotive movement. One most important systems this complex is system for detecting objects on tracks, ruptures bed its turns. Such a can be developed in form camera installed information processing board each rolling stock, or global remote from several locomotives. Regardless implementation there need block images coming cameras. The using interacting full-convolutional convolutional neural networks training dataset covering various situations occurring tracks.