Please use this identifier to cite or link to this item:
https://hdl.handle.net/1889/3817
Full metadata record
DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | Broggi, Alberto | - |
dc.contributor.author | Allodi, Marco | - |
dc.date.accessioned | 2019-04-16T11:29:32Z | - |
dc.date.available | 2019-04-16T11:29:32Z | - |
dc.date.issued | 2019 | - |
dc.identifier.uri | http://hdl.handle.net/1889/3817 | - |
dc.description.abstract | Lane detection is a crucial element for advanced driver assistance systems (ADAS) and fully autonomous driving. In the last decades a lot of progress has been made to realize systems that provides high reliability in every possible scenarios but nowadays most of these systems still work mainly in highways or other highly predictable and structured environments. In this thesis the lane detection problem is studied using approaches based on Convolution Neural Network that represents an extremely powerful framework to understand the context of a scene which is a key requirement for detecting lanes in road images. In particular this thesis focuses on the analysis of the lane detection problem in challenging environments like urban and rural or more generally scenes that present critical lighting, traffic, weather and environmental conditions. Another essential requirement of every perception task for autonomous driving is real-time processing. For this reason the architectures proposed in the thesis are designed to provide the best trade-off between efficiency and accuracy. Regarding the learning procedures, the networks developed are trained to solve an instance segmentation problem to detect the main lane boundaries on a road: ego lane left and right boundaries, left and right lane boundaries. To evaluate the trained models two recently released dataset for lane detection have been used: the TuSimple Lane Detection benchmark, which is composed by images acquired on US highways at daytime, and the BDD100K dataset, which contains road images collected in a wide variety of different environments and conditions. The work performed shows that the implemented architectures and training procedures are able to provide results comparable to other state of the art approaches on the TuSimple Lane Detection Challenge. In the case of more complex and challenging scenarios the presented network models offer very promising results and this is shown with a qualitative comparison with a classic computer vision based lane detection. | it |
dc.language.iso | Inglese | it |
dc.publisher | Università degli studi di Parma. Dipartimento di Ingegneria e architettura | it |
dc.relation.ispartofseries | Dottorato di ricerca in Tecnologie dell'informazione | it |
dc.rights | © Marco Allodi, 2019 | it |
dc.subject | Lane detection | it |
dc.subject | Deep Learning | it |
dc.subject | Convolutional Neural Network | it |
dc.subject | Image segmentation | it |
dc.title | A deep learning approach to lane detection | it |
dc.title.alternative | Un approccio basato su deep learning per l'individuazione delle linee di corsia | it |
dc.type | Doctoral thesis | it |
dc.subject.miur | ING-INF/05 | it |
Appears in Collections: | Tecnologie dell'informazione. Tesi di dottorato |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
relazione_finale_en.pdf Until 2100-01-01 | Relazione finale attività di dottorato | 54.7 kB | Adobe PDF | View/Open Request a copy |
TesiDottorato.pdf | Tesi | 29.61 MB | Adobe PDF | View/Open |
This item is licensed under a Creative Commons License