Author(s):
1. Ivan Ćirić, Univerzitet u Nišu, Mašinski fakultet, Serbia
2. Vlastimir Nikolic, Univerzitet u Nišu, Mašinski fakultet, Serbia
3. Milan Pavlović, College of Applied Technical Sciences Niš, 4. Miloš Simonović, Univerzitet u Nišu, Mašinski fakultet, Serbia
5. Emina Petrovic, Univerzitet u Nišu, Mašinski fakultet, Serbia

Abstract:
In order to increase level of safety on a railway crossings an advanced machine vision system is suggested. Due to different illumination and weather conditions conventional monochromatic and RGB cameras cannot be used as only vision sensor and in this paper authors are suggesting a novel advanced locomotive mounted multi-sensor infra-red camera based system for pedestrian detection on railway crossings. During night and in low-light conditions the suggested systems powered by advanced image processing algorithms can provide reliable pedestrian detection. Advanced image processing algorithms are using artificial intelligence inspired algorithms in order to achieve greater reliability and robustness.

Key words:
image processing, infra-red camera, railway safety, artificial intelligence

Thematic field:
Mechatronics

Date of abstract submission:
10.03.2017.

Conference:
13th International Conference on Accomplishments in Mechanical and Industrial Engineering