OpenVehicleVision database

[yeyang1021]

I want to know , how can I to download the database for the algorithm. I couldn't download them.

[baidut]

I have uploaded all related database to my cloud disk. To download from my cloud disk:

1. open http://yun.baidu.com/share/link?shareid=1552701070&uk=1863129434
2. click 项目公开, then -> OpenVehicleVision and -> datasets

You can also download the database from website

SLD2011

Santaigo Lanes Dataset (SLD2011) website

Related paper： A comparison of gradient versus color and texture analysis for lane detection and tracking

Related video： [Santaigo Lanes Dataset (SLD2011) - Lane changing maneuvers at dusk] youtube Lane tracking with shadows 2 - Bird Eye View - Santaigo Lanes Dataset (SLD2011) youtube

ROMA (ROad MArkings)

website: http://www.lcpc.fr/english/products/image-databases/article/roma-road-markings-1817 related paper: Evaluation of Road Marking Feature Extraction view on IEEE.org

KITTI-ROAD dataset

Road/Lane Detection Evaluation 2013

Please note the license conditions of this software / dataset!

youtube video

• kitti/data_road.zip base kit with: left color images, calibration and training labels (0.5 GB)
• kitti/data_road_right.zip right color image extension (0.5 GB)
• NA grayscale image extension (0.3 GB)
• NA Velodyne laser point extension (1 GB)
• NA OXTS GPS/IMU extension (1 MB)
• kitti/devkit_road.zip development kit (1 MB)
• kitti/devkit_road_mapping.zip Mapping of training set to raw data sequences (1 MB)

[yeyang1021]

thx.

[yeyang1021]

ROMA (ROad MArkings)

website: http://www.lcpc.fr/english/products/image-databases/article/roma-road-markings-1817 related paper: Evaluation of Road Marking Feature Extraction view on IEEE.org

I can not open the url, and I can not find the original image about the ROMA datasets in baiduyun. Only find the groundtruth images. Could you upload the datasets in baiduyun.

[baidut]

Thank you for your feedback.

I have uploaded the raw roma dataset OpenVehicleVision/datasets/ROMA/roma.zip which includes original images (.jpg) and road marking groundtruth (.pgm).

We manually labeled road area ground truth of that dataset, see OpenVehicleVision/datasets/ROMA/ROMA_ROAD_GT.zip

Note ROMA dataset is for research use only and you should cite their paper if you use them. Also, I'd appreciate it if you cite my paper.

@INPROCEEDINGS{Zhenqiang2016,
author={Zhenqiang Ying and Ge Li}, 
booktitle={Acoustics, Speech and Signal Processing (ICASSP), 2016 IEEE International Conference on}, 
title={Robust Lane Marking Detection using Boundary-Based Inverse Perspective Mapping}, 
year={2016}, 
intype = {to appear in},}

Our paper has not published yet, I'd like to give you a copy if you need.

[yeyang1021]

Thank you for your dataset. There are still some problems in the code(Matlab). When run the ICASSP2016.m, the roadDetectionViaBird.m has a error due to the imdump function is missing. You can check the error in the 77 row. Maybe you not provide it. I want to see your paper, if you are convenient. Email: [email protected]

[Spir4L]

Hi, your package is really nice and wel done! By the way I am trying to use your roadDetectionViaBird fucntion on the Kitti dataset but I find that it works badly than on the other ones. Do you have any idea of why?

[baidut]

@Spir4L Thank you for your reply! Kitti dataset is very challenging, I suggest you try FSO:

Project http://abhijitkundu.info/projects/fso/ Code https://bitbucket.org/infinitei/videoparsing

Or use deep learning

Multi-scale context aggregation by dilated convolutions. https://github.com/fyu/dilation

I am still in the holidays, and I will update the code during the school. Sorry to delay your time! Hope it'll help you!

[Spir4L]

@baidut Thank you very much. I will have a look at them. Have a nice day!

[GloriaHM]

hello, I can not download the Velodyne laser point database from the websites. I have tried many times. You cloud disk link is Invalid. Can you update the link again? Thank you for your help.