基于深度学习的驾驶关注区域检测研究与实现

doi:10.16182/j.issn1004731x.joss.18-CVR0694

摘要/Abstract

摘要： 作为智能驾驶的一个关键技术,驾驶关注区域检测方法对智能驾驶或者智能预警系统的性能具有重要影响。针对现有方法存在的不足,基于深度学习,提出一种有效的驾驶关注区域检测方法。采用摄像机自标定的方法得到摄像机的内外参数,利用Canny边缘检测及二分k-means聚类实现消失点的估计,基于所得到的各种估计,建立道路检测模型,利用SSD模型训练得到深度特征,采用SSD卷积层结合FCN8的上采样层进行路面区域检测。在公开数据集上进行了测试,实验结果表明：与现有方法相比,所提的方法不但具有较好的道路检测效果,同时能够比较准确的检测出阴影部分的道路区域。

关键词: 智能驾驶, 驾驶关注区域, 深度学习, 道路检测

Abstract: As a key technology of intelligent driving, driving concern area detection method has an important impact on the performance of intelligent driving or intelligent early warning system. In view of the shortcomings of the existing methods, this paper proposes an effective method for driving concern area detection based on the deep learning. We obtain the camera internal and external parameters by using camera self-calibration method based on camera model, use the Canny edge detection and Bisecting K-means clustering to realize the vanishing point estimation, and establish the road detection model based on the obtained estimates. We obtain the depth features from the SSD model training, use the convolution layer of SSD which combines with the upper sampling layer of FCN8 to detect the region of the road surface. The experimental results show that compared with the existing methods, the proposed method not only has better road detection effect, but also can detect the road area of the shaded part more accurately.

Key words: Intelligent driving, driving concern area, deep learning, road detection

中图分类号:

TP391

叶继华, 时淑霞, 李汉曦, 王仕民, 杨思渝. 基于深度学习的驾驶关注区域检测研究与实现[J]. 系统仿真学报, 2019, 31(7): 1421-1428.

Ye Jihua, Shi Shuxia, Li Hanxi, Wang Shimin, Yang Siyu. Research and Implementation of Driving Concern Area Detection Based on Deep Learning[J]. Journal of System Simulation, 2019, 31(7): 1421-1428.

参考文献

[1] Bertozzi M, Broggi A.GOLD: A parallel real-time stereo vision system for generic obstacle and lane detection[J]. IEEE Transactions on Image Processing (S1057-7149), 1998, 7(1): 62-81.
[2] Kai N, Kezhong H.THMR-V: an effective and robust high-speed system in structured road[C]. 2003 IEEE International Conference on Systems, Man and Cybernetics, 2003. Washington, DC, USA: IEEE, 2003: 4370-4374.
[3] Gregor R, Lutzeler M, Pellkofer M, et al.EMS-Vision: A perceptual system for autonomous vehicles[J]. IEEE Transactions on Intelligent Transportation Systems (S1524-9050), 2002, 3(1): 48-59.
[4] Bertozzi M, Bombini L, Broggi A, et al.GOLD: A framework for developing intelligent-vehicle vision applications[J]. IEEE Intelligent Systems (S1541-1672), 2008, 23(1): 69-71.
[5] Loose H, Franke U.B-spline-based road model for 3d lane recognition[C]. Intelligent Transportation Systems (ITSC), 2010 13^th International IEEE Conference on. Funchal, Madeira Island, Portugal: IEEE, 2010: 91-98.
[6] Benmansour N, Labayrade R, Aubert D, et al.A model driven 3d lane detection system using stereovision[C]. Control, Automation, Robotics and Vision, 2008. ICARCV 2008. 10^th International Conference on. Hanoi, Vietnam: IEEE, 2008: 1277-1282.
[7] Oh C, Kim B, Sohn K.Automatic illumination invariant road detection with stereo vision[C]. Industrial Electronics and Applications (ICIEA), 2012 7th IEEE Conference on. Kota Kinabalu, Malaysia: IEEE, 2012: 889-893.
[8] Thorpe C, Hebert M H, Kanade T, et al.Vision and navigation for the Carnegie-Mellon Navlab[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence (S1939-3539), 1988, 10(3): 362-373.
[9] Guo C, Mita S.A Semantic Graph of Traffic Scenes for Intelligent Vehicle Systems[J]. IEEE Intelligent Systems (S1541-1672), 2012, 27(4): 57-62.
[10] Guo C, Mita S.Semantic-based road environment recognition in mixed traffic for intelligent vehicles and advanced driver assistance systems[C]. Intelligent Transportation Systems (ITSC), 2012 15^th International IEEE Conference on. Bogota, Colombia: IEEE, 2012: 444-450.
[11] Dollár P, Zitnick C L.Structured Forests for Fast Edge Detection[C]. IEEE International Conference on Computer Vision (ICCV). Sydney, Australia: IEEE, 2013: 1841-1848.
[12] Jonathan Long, Evan Shelhamer, Trevor Darrell.Fully Convolutional Networks for Semantic Segmentation[C]. IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Boston, Massachusetts: IEEE, 2015.
[13] Min Sun, Sid Ying-Ze Bao, Silvio Savarese. Object Detection with Geometrical Context Feedback Loop[C]. British Machine Vision Conference (BMVC). Aberystwyth, UK: IEEE, 2010.
[14] Min Sun, Gary Bradski, Bing-Xin Xu, et al.Depth-encoded Hough Voting for Coherent Object Detection, Pose Estimation, and Shape Recovery[C]. European Conference on Computer Vision (ECCV). Crete, Greece:Springer, 2010.
[15] Zhang Z.A flexible new technique for camera calibration[J]. IEEE Trans Pattern Analysis and Machine Intelligence (S1939-3539), 2000, 22(11): 1330-1334.
[16] Li Y T, Zhu F H, Ai Y F, et al.On automatic and dynamic camera calibration based on traffic visual surveillance[C]. Intelligent Vehicles Symposium, 2007 IEEE. Istanbul, Turkey: IEEE, 2007: 358-363.
[17] Wang K, Huang H, Li Y, et al.Research on Lane-Marking line based camera calibration[C]. IEEE International Conference on Vehicular Electronics and Safety. Beijing, china: IEEE, 2007: 1-6.
[18] Moghadam P, Starzyk J A.Fast vanishing-point detection algorithm in unstructured environments[J]. IEEE Trans. Image Process (S1057-7149), 2012, 21(1): 425-430.
[19] Ebrahimpour R, Rasoolinezhad R, Hajiabolhasni Z, et al.Vanishing point detection in corridors: using Hough transform and K-means clustering[J]. IET Comput. Vis.(S1751-9632), 2012, 6(1): 40-51.
[20] Kosecká J, Zhang W.Video compass[C]. Proc. Seventh European Conf. on Computer Vision. Copenhagen: Springer, 2002: 476-490.
[21] Weili D, Yong L, Wenfeng W, et al.Vanishing point detection algorithm for urban road image based on the envelope of perpendicular and parallel lines[J]. Acta Optica Sinica (S0253-2239), 2014, 34(10): 1015002.
[22] Wei W, Zhijing Y, Yin G, et al.Study on the occlusion problem in dynamic space intersection measurement with multi-camera systems[J]. Acta Optica Sinica (S0253-2239), 2014, 34(4): 0415003.
[23] Marvin Teichmann, Michael Weber, Marius Zöllner, et al.MultiNet: Real-time Joint Semantic Reasoning for Autonomous Driving[C]. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, Nevada, USA: IEEE, 2016.
[24] https://github.com/NVIDIA/DIGITS/blob/v4.0.0-rc.3/digits/extensions/data/objectDetection/README.md
[25] 梁军. 基于Multi-Agent和驾驶行为的汽车追尾预警系统关键理论与技术研究[D]. 苏州: 江苏大学, 2015.
Liang Jun.Research on the Key Theory and Technology of Car Rear-End Collision Alarm System Based on Multi-Agent and Driving Behavior[D]. Shuzhou: Jiangsu University, 2015.
[26] Ankit Laddha, Mehmet Kemal Kocamaz, Luis E. Navarro-Serment,Martial Hebert.Map-Supervised Road Detection[C]. IEEE Intelligent Vehicles Symposium (IV). Gothenburg,Sweden: IEEE, 2016: 118-123.
[27] Hojato lah Amimi, Babak Karasfi. New Approach to Road Detection in Challenging Outdoor Environment for Autonomous Vehicle[C]. Artificial Intelligence and Robotics. Qazvin, Iran: IEEE, 2016: 7-11.