摘要
随着信息技术的不断发展,无人驾驶技术得到了国内外的广泛研究,但是其核心技术能否得到突破取决于无人驾驶系统的智能化水平,图像识别技术在其中充当了很重要的角色。为提升交通场景下图像的语义分割精度,文章开展了基于深度学习网络的图像语义分割算法研究。首先文章研究了基于窗口局部算法的版权局立体匹配算法,得到了视图差D;然后将其余RGB格式的图像融合,形成RGB-D模式的图像,以此构件了文章的图像数据库,并基于图像数据库,选择了不同的深度学习网络架构及学习策略,对网络进行了训练。最后通过对训练结果的对比分析,得出了以下结论:当选择相同的网络架构和相同的学习策略的前提下,采用文章所提出的基于RGB-D格式作为图像输入,在全局精确度和平均精确度方面表现更好。
With the rapid development of information technology, unmanned driving technology has been widely studied at home and abroad, but whether its core technology can be broken through depends on the intelligent level of unmanned driving system. Image recognition technology plays an important role in it. In order to improve the accuracy of image semantics segmentation in traffic scenes, this paper studies image semantics segmentation algorithm based on deep learning network. Firstly, the stereo matching algorithm of Copyright Bureau based on window local algorithm is studied, and the view difference D is obtained. Then, the images of other RGB formats are fused to form RGB-D mode images, which constitute the image database of this paper. Based on the image database, different deep learning network architectures and learning strategies are selected to train the network. Finally, through the comparative analysis of the training results, the following conclusions are drawn: under the premise of choosing the same network architecture and the same learning strategy, using the RGB-D format proposed in this paper as the image input, the global accuracy and average accuracy are better.
引文
[1]Brown B G,Murphy A H.Improving forecasting performance by combining forecasts:The example of road‐surface temperature forecasts[J].Meteorological Applications,2010,3(3):257-265.
[2]Chen K,Lu M,Tan G,et al.CRSM:Crowdsourcing Based Road Surface Monitoring[C]// IEEE International Conference on High Performance Computing & Communications & IEEE International Conference on Embedded & Ubiquitous Computing.2014.
[3]高常鑫,桑农.基于深度学习的高分辨率遥感影像目标检测[J].测绘通报,2014(s1):108-111.
[4]高凯珺,孙韶媛,姚广顺,等.基于深度学习的无人车夜视图像语义分割[J].应用光学,2017,38(3):421-428.
[5]刘丹,刘学军,王美珍.一种多尺度CNN 的图像语义分割算法[J].遥感信息,2017,32(1):57-64.
[6]KRIZHEVSKY A,SUTSKEVER I,HINTON G E.Image Net classification with deep convolutional neural networks[J].Advances in Neural Information Processing Systems,2012,25(2):1-9.
[7]SIMONYAN K,ZISSERMAN A.Very deep convolutional networks for large-scale image recognition[J].ArXiv Preprint,ArXiv:1409.1556,2014.
[8]SZEGEDY C,LIU W,JIA Y,et al.Going deeper with convolutions[C]//IEEE Conference on Computer Vision and Pattern Recognition,2014:1-9.
[9]DENG J,DONG W,SOCHER R,et al.ImageNet:a large-scale hierarchical image database[C]// IEEE Computer Vision and Pattern Recognition,2009:248-255.
[10]BADRINARAYANAN V,HANDA A,CIPOLLA R.SegNet:a deep convolutional encoder-decoder architecture for robust semantic pixel-wise labelling[J].ArXiv Preprint,ArXiv,1505,07293,2015.
[11]BADRINARAYANAN V,KENDALL A,CIPOLLA R.SegNet:a deep convolutional encoder-decoder architecture for image segmentation[J].IEEE Transactions on Pattern Analysis & Machine Intelligence,2017,PP(99):1.
[12]NOH H,HONG S,HAN B.Learning deconvolution network for semantic segmentation[C]//IEEE International Conference on Computer Vision,2015:1520-1528.
[13]GUPTA S,GIRSHICK R,ARBELáEZ P,et al.Learning rich features from RGB-D images for object detection and segmentation[C]// European Conference on Computer Vision.2014:345-360.
[14]FILLIAT D,BATTESTI E,BAZEILLE S,et al.RGBD object recognition and visual texture classification for indoor semantic mapping[C]//2012 IEEE International Conference on Technologies for Practical Robot Applications.2012:127-132.
[15]LI L,HUANG H,QIAN B,et al.Vehicle detection method based on mean shift clustering[J].Journal of Intelligent & Fuzzy Systems,2016,31(3):1355-1363.