Detection of Traffic Blackspots Using Deep Learning for Autonomous Vehicles with Street View Imagery
DOI:
https://doi.org/10.37266/ISER.2022v10i1.pp34-43Keywords:
Traffic Blackspots, Autonomous Vehicles, Deep Learning, Convolutional Neural Networks, Image Understanding, Image SegmentationAbstract
Accidents have been a major cause of fatalities and injuries, much of which is attributed to human errors such as over speeding and drunk-driving. The onset of Autonomous Vehicles delivers a promising future with accident rate reducing significantly. But we also cannot deny the fact that these systems struggle in certain cases to avoid accidents. Hence, we propose a method to alert these autonomous vehicles in advance that they are approaching a Traffic Blackspot. A Traffic Blackspot is a dangerous prone-to-accident spot or location. This approach leverages the power of Deep Learning to understand the environmental factors of each location through street view images. This automated system will have access to the GPS location of the autonomous vehicle which it matches with its database of blackspots to determine if the vehicle is in or around such a location and alert the system when it is in the vicinity of a Blackspot. Due to this alert, a timely control can be adopted by the system with respect to speed, position on the road, etc., thus reducing the possibility of an autonomous vehicle potentially meeting with an accident.
References
Chilamkurthy, Sasank (2017). A 2017 Guide to Semantic Segmentation with Deep Learning. Retrieved from: https://blog.qure.ai/notes/semantic-segmentation-deep-learning-review
Daimler AG, MPI Informatics, and TU Darmstadt (2020). Cityscapes Dataset. Retrieved from: https://www.cityscapes-dataset.com/
Eliot, Lance (2019). Essential Stats For Justifying And Comparing Self-Driving Cars To Humans At The Wheel. Retrieved from: https://www.forbes.com/sites/lanceeliot/2019/05/30/essential-stats-for-justifying-and-comparing-self-driving-cars-to-humans-at-the-wheel/?sh=407b671346ed
Caner, Filiz (2020). Can Autonomous Vehicles Prevent Traffic Accidents? Retrieved from: https://www.intechopen.com/books/accident-analysis-and-prevention/can-autonomous-vehicles-prevent-traffic-accidents-
Hoffman, Chris (2017). How to See Exactly Where a Photo Was Taken (and Keep Your Location Private). Retrieved from: https://www.howtogeek.com/211427/how-to-see-exactly-where-a-photo-was-taken-and-keep-your-location-private/#:~:text=Modern%20smartphones%20
Hyeoksin-ro, Wonju-si, Gangwon-do, and Korea KoROAD (2020). What causes traffic accidents? Retrieved from: https://www.koroad.or.kr/en_web/view/trfEnv4.do
Long, Jonathan, Shelhamer, Evan, and Darrell, Trevor (2015, March). Fully Convolutional Networks for Semantic Segmentation. Retrieved from: https://arxiv.org/pdf/1411.4038.pdf
Rahiman, Wan (2013). An overview of development GPS navigation for autonomous car. Retrieved from: https://www.researchgate.net/publication/261088859_An_overview_of_development_GPS_navigation_for_autonomous_car
Saha, Sumit (2018). A Comprehensive Guide to Convolutional Neural Networks — the ELI5 way. Retrieved from: https://towardsdatascience.com/a-comprehensive-guide-to-convolutional-neural-networks-the-eli5-way-3bd2b1164a53
Several authors (2020). ArcGIS API for Python; How DeepLab V3 Works. Retrieved from: https://developers.arcgis.com/python/guide/how-deeplabv3-works/
Sharma, Pulkit (2019). Computer Vision Tutorial: A Step-by-Step Introduction to Image Segmentation Techniques (Part 1). Retrieved from: https://www.analyticsvidhya.com/blog/2019/04/introduction-image-segmentation-techniques-python/
Sik-Ho, Tsang (2018). Review: FCN — Fully Convolutional Network (Semantic Segmentation). Retrieved from: https://towardsdatascience.com/review-fcn-semantic-segmentation-eb8c9b50d2d1
Sik-Ho, Tsang (2019). Review: DeepLabv3 — Atrous Convolution (Semantic Segmentation). Retrieved from: https://towardsdatascience.com/review-deeplabv3-atrous-convolution-semantic-segmentation-6d818bfd1d74
Statista Research Department (2021, March). Number of deaths due to road accidents in India 2005-2019. Retrieved from: https://www.statista.com/statistics/746887/india-number-of-fatalities-in-road-accidents/
Tan, Ying (2016, May). Gpu-Based Parallel Implementation of Swarm Intelligence Algorithms. Retrieved from: https://www.sciencedirect.com/science/article/pii/B978012809362750011X
Tiu, Ekin (2019, August). Metrics to Evaluate your Semantic Segmentation Model. Retrieved from: https://towardsdatascience.com/metrics-to-evaluate-your-semantic-segmentation-model-6bcb99639aa2
Young, Joe (2020, June). Self-driving vehicles could struggle to eliminate most crashes. Retrieved from: https://www.iihs.org/news/detail/self-driving-vehicles-could-struggle-to-eliminate-most-crashes
Zhao, Jianfeng, Liang Bodong, and Chen Qiuxia (2018, January). The key technology toward the self-driving car. Retrieved from: https://www.emerald.com/insight/content/doi/10.1108/IJIUS-08-2017-0008/full/html
Published
How to Cite
Issue
Section
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
The copyediting stage is intended to improve the flow, clarity, grammar, wording, and formatting of the article. It represents the last chance for the author to make any substantial changes to the text because the next stage is restricted to typos and formatting corrections. The file to be copyedited is in Word or .rtf format and therefore can easily be edited as a word processing document. The set of instructions displayed here proposes two approaches to copyediting. One is based on Microsoft Word's Track Changes feature and requires that the copy editor, editor, and author have access to this program. A second system, which is software independent, has been borrowed, with permission, from the Harvard Educational Review. The journal editor is in a position to modify these instructions, so suggestions can be made to improve the process for this journal.
