Road Recognition and Lane Detection using Deep Learning

IF 5.8 3区管理学 Q1 COMPUTER SCIENCE, INFORMATION SYSTEMS

Journal of Information Technology Pub Date : 2023-07-31 DOI:10.59185/dg8xmp38

Md. Fazlul Karim Patwary, Moumita Chanda, Sadiya Rahman, Md. Tanvir Ahmed

{"title":"Road Recognition and Lane Detection using Deep Learning","authors":"Md. Fazlul Karim Patwary, Moumita Chanda, Sadiya Rahman, Md. Tanvir Ahmed","doi":"10.59185/dg8xmp38","DOIUrl":null,"url":null,"abstract":"An autonomous vehicle needs to be familiar with its surroundings. The safety of thetransportation system is greatly enhanced by advanced driving assistance systems (ADASs). Road detectionis one of the steps that a driving car must do. Is it possible for a computer to recognize a road in a singlephotograph for this purpose? This question is addressed using the lane detecting techniques. Roads andlanes are tough for machine learning to differentiate because of training a machine to recognize a road.Over the past few decades, a number of lane identification technologies have been created and integratedinto various autonomous cars. It is still very difficult to create lane recognition technology that caneffectively identify a road lane in a range of road conditions. This research provides a composite approachfor road detection from image processing using convolutional neural networks by testing 150 photographsthat include a road, jungle, muddy road, and barriers. It will decide if an image contains a road or not. Inthis essay, we first establish whether a road exists. The second step is to find a lane on the finished road.The benefit of the proposed technology is that if there is a road, the automobile can continue to moveforward; otherwise, it will stop.","PeriodicalId":50178,"journal":{"name":"Journal of Information Technology","volume":null,"pages":null},"PeriodicalIF":5.8000,"publicationDate":"2023-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Information Technology","FirstCategoryId":"91","ListUrlMain":"https://doi.org/10.59185/dg8xmp38","RegionNum":3,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"COMPUTER SCIENCE, INFORMATION SYSTEMS","Score":null,"Total":0}

引用次数: 0

Abstract

An autonomous vehicle needs to be familiar with its surroundings. The safety of thetransportation system is greatly enhanced by advanced driving assistance systems (ADASs). Road detectionis one of the steps that a driving car must do. Is it possible for a computer to recognize a road in a singlephotograph for this purpose? This question is addressed using the lane detecting techniques. Roads andlanes are tough for machine learning to differentiate because of training a machine to recognize a road.Over the past few decades, a number of lane identification technologies have been created and integratedinto various autonomous cars. It is still very difficult to create lane recognition technology that caneffectively identify a road lane in a range of road conditions. This research provides a composite approachfor road detection from image processing using convolutional neural networks by testing 150 photographsthat include a road, jungle, muddy road, and barriers. It will decide if an image contains a road or not. Inthis essay, we first establish whether a road exists. The second step is to find a lane on the finished road.The benefit of the proposed technology is that if there is a road, the automobile can continue to moveforward; otherwise, it will stop.

查看原文本刊更多论文

利用深度学习进行道路识别和车道检测

自动驾驶汽车需要熟悉周围环境。先进的驾驶辅助系统（ADAS）大大提高了交通系统的安全性。道路检测是自动驾驶汽车必须完成的步骤之一。为此，计算机是否有可能在单张照片中识别道路？这个问题可以通过车道检测技术来解决。道路和车道对于机器学习来说是很难区分的，因为要训练机器识别道路。在过去的几十年里，人们创造了许多车道识别技术，并将其集成到各种自动驾驶汽车中。在过去的几十年里，人们创造了许多车道识别技术，并将其集成到各种自动驾驶汽车中。但要创造出在各种路况下都能有效识别车道的车道识别技术，仍然非常困难。本研究通过测试 150 张照片，包括道路、丛林、泥泞道路和障碍物，提供了一种利用卷积神经网络从图像处理中检测道路的复合方法。它将决定一幅图像是否包含道路。在本文中，我们首先确定道路是否存在。这项技术的好处在于，如果有路，汽车就能继续前行；反之，汽车就会停下。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Journal of Information Technology 工程技术-计算机：信息系统

CiteScore

10.00

自引率

1.80%

发文量

审稿时长

>12 weeks

期刊介绍： The aim of the Journal of Information Technology (JIT) is to provide academically robust papers, research, critical reviews and opinions on the organisational, social and management issues associated with significant information-based technologies. It is designed to be read by academics, scholars, advanced students, reflective practitioners, and those seeking an update on current experience and future prospects in relation to contemporary information and communications technology themes. JIT focuses on new research addressing technology and the management of IT, including strategy, change, infrastructure, human resources, sourcing, system development and implementation, communications, technology developments, technology futures, national policies and standards. It also publishes articles that advance our understanding and application of research approaches and methods.