基于编解码器模型的车道识别与车辆检测算法
|
摘要
针对无人驾驶车辆环境感知问题,通过编码器提取共享图像特征,再通过解码器来实现语义分割、分类和目标检测模块,并应用在车道识别和车辆检测上.在无人驾驶中,任务的实时性非常关键,这种共享编码器模型能一定程度上提高任务实时性.实验结果表明,该模型的语义分割在KITTI数据集上的平均精度达到93.89%,比最优性能提升0.53%,联合检测速度达到25.43 Hz.
Aiming at the problem of environment perception of self-driving, this paper semantic segmentation,classification and target detection module are realized by the code model, which is applied to lane recognition and vehicle detection. Shared image features are extracted by encoder, and three different functions are realized by decoder. This Shared encoder model can improve the real-time performance of tasks. In self-driving, the real-time performance of tasks is the key. Experimental results show that the average precision of semantic segmentation of this model on KITTI dataset reaches 93.89%, which is 0.53% higher than the optimal performance, and the joint detection speed reaches 25.43 Hz.
Aiming at the problem of environment perception of self-driving, this paper semantic segmentation,classification and target detection module are realized by the code model, which is applied to lane recognition and vehicle detection. Shared image features are extracted by encoder, and three different functions are realized by decoder. This Shared encoder model can improve the real-time performance of tasks. In self-driving, the real-time performance of tasks is the key. Experimental results show that the average precision of semantic segmentation of this model on KITTI dataset reaches 93.89%, which is 0.53% higher than the optimal performance, and the joint detection speed reaches 25.43 Hz.
引文
[1]KRIZHEVSKY A,SUTSKEVE R I,HINTONN G E.Imagenet classification with deep convolutional neural networks[J].Advances in Neural Information Processing Systems,2017,60(6):84-90.
[2]WU Z,SHEN C,HENGEL A.Wider or deeper:revisiting the resnet model for visual recognition[J].arXiv preprint arXiv:1611.10080,2016.
[3]HARIHARAN B,ARBELA Z P,GIRSHICK R,et al.Simultaneous detection and segmentation[C]//European Conference on Computer Vision.Zurich,Switzerland:Springer,2014:297-312.
[4]郭继舜.面向自动驾驶的语义分割和目标检测技术[D].成都:电子科技大学,2018.
[5]HOLSCHNEIDER M,KRONLANDMARTINET R,MOR-LET J,et al.A real-time algorithm for signal analysis with the help of the wavelet transform[C]//Wavelets.Berlin,Heidelberg:Springer,1990:286-297.
[6]RUSSAKOVSK,YUSSAKOVSKY O,DENG J,et al.Imagenet large scale visual recognition challenge[J].International Journal of Computer Vision,2015,115(3):211-252.
[7]HE K,ZHANG X,REN S,et al.Deep residual learning for image recognition[C]//Proceedings of the IEEE conference on computer vision and pattern recognition.In las Vegas:IEEE,2016:770-778.
[8]陈良甫,杨曾.一种基于残差网络的多任务模型[J].中国集成电路,2017(8):64-71.CHEN L F,YANG Z.A residual network based multi-task model[J].China Integrated Circuit,2017(8):64-71.
[9]MA W C,WANG S,BRUBAKER M A,et al.Find your way by observing the sun and other semantic cues[C]//Robotics and Automation(ICRA),2017 IEEE International Conference on.Singapore:IEEE,2017:6292-6299.
[10]GIUSTI A,CIRE?AN D C,MASCI J,et al.Fast image scanning with deep max-pooling convolutional neural networks[C]//2013 IEEE International Conference on Image Processing.Melbourne,Australia:IEEE,2013:4034-4038.
[11]LI H,ZHAO R,WANG X.Highly efficient forward and backward propagation of convolutional neural networks for pixelwise classification[J].arXiv preprint arXiv:1412.4526,2014.
[12]LONG J,SHELHAMER E,DARRELL T.Fully convolutional networks for semantic segmentation[C]//Proceedings of the IEEE conference on computer vision and pattern recognition.Massachusetts:IEEE,2015:3431-3440.
[13]ZEILER M D,KRISHNAN D,TAYLOR G W,et al.Deconvolutional networks[C]//Computer Vision&Pattern Recognition.San Francisco:IEEE Computer Society,2010.
[14]HE K,ZHANG X,REN S,et al.Spatial pyramid pooling indeep convolutional networks for visual recognition[J].IEEEtransactions on pattern analysis and machine intelligence,2015,37(9):1904-1916.
[15]HOSANG J,BENENSON R,DOLLáR P,et al.What makes for effective detection proposals?[J].IEEE transactions on pattern analysis and machine intelligence,2016,38(4):814-830.
[16]Ren S Q,He K M,GIRSHICK R,et al.Faster r-cnn:towards real-time object detection with region proposal networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(6):1137-1149.
[17]ERHAN D,SZEGEDY C,TOSHEV A,et al.Scalable object detection using deep neural networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Canada:NIPS,2014:2147-2154.
[18]LI H,ZHAO R,WANG X.Highly efficient forward and backward propagation of convolutional neural networks for pixelwise classification[J].arXiv preprint arXiv:1412.4526,2014.
[19]SERMANET P,EIGEN D,ZHANG X,et al.Overfeat:integrated recognition,localization and detection using convolutional networks[J].arXiv preprint arXiv:1312.6229,2013.
[20]GEIGER A,LENZ P,STILLER C,et al.Vision meets robotics:The KITTI dataset[J].The International Journal of Robotics Research,2013,32(11):1231-1237.
[21]FRITSCH J,KUEHNL T,GEIGER A.A New Performance Measure and Evaluation Benchmark for Road Detection Algorithms[C]//IEEE Int.Conf.on Intelligent Transportation Systems(ITSC).The Hague:IEEE,2013.
[22]URTASUN R,LENZ P,GEIGER A.Are we ready for autonomous driving?The KITTI vision benchmark suite[C]//2012 IEEE Conference on Computer Vision and Pattern Recognition.Washington:IEEE Computer Society,2012.
[23]KINGMA D P,BA J.Adam:a method for stochastic optimization[J].arXiv preprint arXiv:1412.6980,2014.
[2]WU Z,SHEN C,HENGEL A.Wider or deeper:revisiting the resnet model for visual recognition[J].arXiv preprint arXiv:1611.10080,2016.
[3]HARIHARAN B,ARBELA Z P,GIRSHICK R,et al.Simultaneous detection and segmentation[C]//European Conference on Computer Vision.Zurich,Switzerland:Springer,2014:297-312.
[4]郭继舜.面向自动驾驶的语义分割和目标检测技术[D].成都:电子科技大学,2018.
[5]HOLSCHNEIDER M,KRONLANDMARTINET R,MOR-LET J,et al.A real-time algorithm for signal analysis with the help of the wavelet transform[C]//Wavelets.Berlin,Heidelberg:Springer,1990:286-297.
[6]RUSSAKOVSK,YUSSAKOVSKY O,DENG J,et al.Imagenet large scale visual recognition challenge[J].International Journal of Computer Vision,2015,115(3):211-252.
[7]HE K,ZHANG X,REN S,et al.Deep residual learning for image recognition[C]//Proceedings of the IEEE conference on computer vision and pattern recognition.In las Vegas:IEEE,2016:770-778.
[8]陈良甫,杨曾.一种基于残差网络的多任务模型[J].中国集成电路,2017(8):64-71.CHEN L F,YANG Z.A residual network based multi-task model[J].China Integrated Circuit,2017(8):64-71.
[9]MA W C,WANG S,BRUBAKER M A,et al.Find your way by observing the sun and other semantic cues[C]//Robotics and Automation(ICRA),2017 IEEE International Conference on.Singapore:IEEE,2017:6292-6299.
[10]GIUSTI A,CIRE?AN D C,MASCI J,et al.Fast image scanning with deep max-pooling convolutional neural networks[C]//2013 IEEE International Conference on Image Processing.Melbourne,Australia:IEEE,2013:4034-4038.
[11]LI H,ZHAO R,WANG X.Highly efficient forward and backward propagation of convolutional neural networks for pixelwise classification[J].arXiv preprint arXiv:1412.4526,2014.
[12]LONG J,SHELHAMER E,DARRELL T.Fully convolutional networks for semantic segmentation[C]//Proceedings of the IEEE conference on computer vision and pattern recognition.Massachusetts:IEEE,2015:3431-3440.
[13]ZEILER M D,KRISHNAN D,TAYLOR G W,et al.Deconvolutional networks[C]//Computer Vision&Pattern Recognition.San Francisco:IEEE Computer Society,2010.
[14]HE K,ZHANG X,REN S,et al.Spatial pyramid pooling indeep convolutional networks for visual recognition[J].IEEEtransactions on pattern analysis and machine intelligence,2015,37(9):1904-1916.
[15]HOSANG J,BENENSON R,DOLLáR P,et al.What makes for effective detection proposals?[J].IEEE transactions on pattern analysis and machine intelligence,2016,38(4):814-830.
[16]Ren S Q,He K M,GIRSHICK R,et al.Faster r-cnn:towards real-time object detection with region proposal networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(6):1137-1149.
[17]ERHAN D,SZEGEDY C,TOSHEV A,et al.Scalable object detection using deep neural networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Canada:NIPS,2014:2147-2154.
[18]LI H,ZHAO R,WANG X.Highly efficient forward and backward propagation of convolutional neural networks for pixelwise classification[J].arXiv preprint arXiv:1412.4526,2014.
[19]SERMANET P,EIGEN D,ZHANG X,et al.Overfeat:integrated recognition,localization and detection using convolutional networks[J].arXiv preprint arXiv:1312.6229,2013.
[20]GEIGER A,LENZ P,STILLER C,et al.Vision meets robotics:The KITTI dataset[J].The International Journal of Robotics Research,2013,32(11):1231-1237.
[21]FRITSCH J,KUEHNL T,GEIGER A.A New Performance Measure and Evaluation Benchmark for Road Detection Algorithms[C]//IEEE Int.Conf.on Intelligent Transportation Systems(ITSC).The Hague:IEEE,2013.
[22]URTASUN R,LENZ P,GEIGER A.Are we ready for autonomous driving?The KITTI vision benchmark suite[C]//2012 IEEE Conference on Computer Vision and Pattern Recognition.Washington:IEEE Computer Society,2012.
[23]KINGMA D P,BA J.Adam:a method for stochastic optimization[J].arXiv preprint arXiv:1412.6980,2014.