Powered by OpenAIRE graph
Found an issue? Give us feedback
addClaim

This Research product is the result of merged Research products in OpenAIRE.

You have already added 0 works in your ORCID record related to the merged Research product.

Разработка алгоритмов для Ð±ÐµÑÐ¿Ð¸Ð»Ð¾Ñ‚Ð½Ñ‹Ñ Ñ‚Ñ€Ð°Ð½ÑÐ¿Ð¾Ñ€Ñ‚Ð½Ñ‹Ñ ÑÑ€ÐµÐ´ÑÑ‚Ð²

выпускная квалификационная работа бакалавра

Разработка алгоритмов для Ð±ÐµÑÐ¿Ð¸Ð»Ð¾Ñ‚Ð½Ñ‹Ñ Ñ‚Ñ€Ð°Ð½ÑÐ¿Ð¾Ñ€Ñ‚Ð½Ñ‹Ñ ÑÑ€ÐµÐ´ÑÑ‚Ð²

Abstract

В данной работе рассматривается использование алгоритмов компьютерного зрения и машинного обучения для их применения в беспилотных транспортных средствах, в частности беспилотный автомобилях. Более подробно, выделяется внимание на задачу обработки изображении из видеопотока для получения информации об окружаюшего пространства, способствуя возможности дальнейшего принятия решения об управляемых действиях автомобиля. Целью работы является разработка системы зрения для беспилотного автомобиля, позволяющая воспринять окружающее пространство из видеопотока. Данная система на входе должна принимать кадры из видеопотока, при этом предполагается что камера находится в передней части автомобиля и направлена вперед, глядя на пространство перед автомобилем. В качестве выхода система должна отдавать размеченные данные дороги. Дополнительно рассмотрена задача построения траектория движения при анализе размеченных данных дороги. При выполнения этой работы были рассмотрены различные способы обработки изображения, исходя из алгоритмов компьютерного зрения и машинного обучения. Проверка работы и оценка качество алгоритмов проводилось на имеющихся исходных видео файлов с высoким разрешением со съемкой гоночной трассы. Результирующие метки качественно выделяют дорогу даже при плохих условиях и являются достаточными для использования в дальнейших этапах систем зрения в беспилотных автомобилях.

This paper discusses the use of computer vision and machine learning algorithms for their use in autonomous vehicles, in particular self-driving cars. In more detail, the work is focused on the task of processing images from a video stream to obtain information about the surrounding environment, contributing to the possibility of further decision-making for controlling actions of the vehicle. The aim of the work is to develop a vision system for an autonomous vehicle, which allows to perceive the surrounding environment from a video stream. This system should receive frames from a video stream as input, assuming that the camera is in front of the vehicle and is directed forward, looking at the space in front of the vehicle. As an output, the system should return labeled road data. In addition, the problem of constructing a movement trajectory from analysis of the labeled road data is considered. In carrying out this work, various methods of image processing were considered based on computer vision and machine learning algorithms. Verification of the work and evaluation of the quality of the algorithms was carried out on the available source video files with high resolution taken on a racing track. The resulting labels extract the road with high quality even under poor conditions and are sufficient for use in subsequent stages of vision systems in autonomous vehicles.

Keywords

vision system, autonomous vehicle, система зрения, беспилотный автомобиль, семантическая сегментация, semantic segmentation, полносверточные нейронные сети, fully convolutional neural networks

  • BIP!
    Impact byBIP!
    selected citations
    These citations are derived from selected sources.
    This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    0
    popularity
    This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
    Average
    influence
    This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    Average
    impulse
    This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
    Average
Powered by OpenAIRE graph
Found an issue? Give us feedback
selected citations
These citations are derived from selected sources.
This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Citations provided by BIP!
popularity
This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
BIP!Popularity provided by BIP!
influence
This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Influence provided by BIP!
impulse
This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
BIP!Impulse provided by BIP!
0
Average
Average
Average
Upload OA version
Are you the author of this publication? Upload your Open Access version to Zenodo!
It’s fast and easy, just two clicks!