Ways for Improving Efficiency of Computer Vision for Autonomous Vehicles and Driver Assistance Systems

This paper presents analysis of human vision features which affects the development of systems of computer vision. Human eyes have known limitations and defects. One of characteristics of vision, for example, the field of vision should be taken into account at the arrangement of cameras on an unmanned vehicle. This characteristic has a great influence on safety of driving and can considerably change depending on many factors. The authors analyzed the range and reasons of changes of the mentioned factor. Also, the paper presents parameters for assessment of the field of vision. Computer vision systems have to fulfill many requirements connected with operational conditions, for example, backlight, illumination overfalls, and great changes of brightness, motion in darkness. The variant of arrangement of two cameras for proving binocular vision is presented.