基于栈式卷积自编码的视觉SLAM闭环检测
|
摘要
同时定位与构图(SLAM)主要用于解决移动机器人在未知环境中进行地图构建和导航的问题,是移动机器人实现自主移动的基础.闭环检测是视觉SLAM的关键步骤,对构建一致性地图和减少位姿累积误差具有重要作用.当前的闭环检测方法通常采用传统的SIFT、SURF等特征,很容易受到环境影响,为了提高闭环检测的准确性和鲁棒性,提出基于无监督栈式卷积自编码(CAEs)模型的特征提取方法,运用训练好的CAEs卷积神经网络对输入图像进行学习,将输出的特征应用于闭环检测.实验结果表明:与传统的BoW方法及其他基于深度学习模型的方法相比,所提出的算法能够有效降低图像特征的维数并改善特征描述的效果,可以在机器人SLAM闭环检测环节获得更好的精确性和鲁棒性.
As the foundation to realize the autonomous movement of mobile robots, simultaneous localization and mapping(SLAM), which is mainly used to solve the problem of mobile robots mapping and navigation in unknown environment, has been paid more attention in recent years. Loop closure detection, one of the key steps of visual SLAM,plays an important role to make a globally consistent map and reduce accumulated error of robot pose. Current methods for loop closure detection are vulnerable to environmental influence because they always adopt traditional features such as SIFT and SURF. To improve the accuracy and robustness of loop closure detection, a method based on unsupervised Stacked Convolutional Autoencoders(CAEs) model is proposed. The trained CAEs convolution neural network is used to learn from input images, while the output features are used for loop closure detection. The results of experiment show that the proposed method, compared with traditional BoW-based methods and other methods based on deep learning model, can effectively reduce the dimension of image features and improve the effect of feature description. Thus, it can attain better accuracy and robustness in loop closure detection of robot SLAM.
As the foundation to realize the autonomous movement of mobile robots, simultaneous localization and mapping(SLAM), which is mainly used to solve the problem of mobile robots mapping and navigation in unknown environment, has been paid more attention in recent years. Loop closure detection, one of the key steps of visual SLAM,plays an important role to make a globally consistent map and reduce accumulated error of robot pose. Current methods for loop closure detection are vulnerable to environmental influence because they always adopt traditional features such as SIFT and SURF. To improve the accuracy and robustness of loop closure detection, a method based on unsupervised Stacked Convolutional Autoencoders(CAEs) model is proposed. The trained CAEs convolution neural network is used to learn from input images, while the output features are used for loop closure detection. The results of experiment show that the proposed method, compared with traditional BoW-based methods and other methods based on deep learning model, can effectively reduce the dimension of image features and improve the effect of feature description. Thus, it can attain better accuracy and robustness in loop closure detection of robot SLAM.
引文
[1]张亮,蒋荣欣,陈耀武.移动机器人在未知环境下的同步定位与地图重建方法[J].控制与决策, 2010, 25(4):515-520.(Zhang L, Jiang R X, Chen Y W. An improved fast SLAM algorithm for mobile robots’ simultaneous localization and mapping in unknown environments[J]. Control and Decision, 2010, 25(4):515-520.)
[2] Williams B, Klein G, Reid I. Automatic relocalization and loop closing for real-time monocular SLAM[J]. IEEE Trans on Pattern Analysis and Machine Intelligence,2011, 33(9):1699-1712.
[3] Kawewong A, Tongprasit N, Hasegawa O. A speeded-up online incremental vision-based loop-closure detection for long-term SLAM[J]. Advanced Robotics, 2013,27(17):1325-1336.
[4]赵洋,刘国良,田国会,等.基于深度学习的视觉SLAM综述[J].机器人, 2017, 39(6):889-896.(Zhao Y, Liu G L, Tian G H, et al. A survey of visual SLAM based on deep learning. robot[J]. Robot, 2017,39(6):889-896.)
[5] Ng P C, Henikoff S. SIFT:Predicting amino acid changes that affect protein function[J]. Nucleic Acids Research,2003, 31(13):3812-3814.
[6] Bay H, Ess A, Tuytelaars T, et al. Speeded-up robust features(SURF)[J]. Computer Vision and Image Understanding, 2008, 110(3):346-359.
[7] Shekhar R, Jawahar C V. Word image retrieval using bag of visual words[C]. IEEE 10th IAPR Int Workshop on Document Analysis Systems(DAS). Gold Coast:IEEE,2012:297-301.
[8] Krizhevsky A, Sutskever I, Hinton G E. Imagenet classification with deep convolutional neural networks[C]. Advances in neural information processing systems. Lake Tahoe:Curran Associates Inc, 2012:1097-1105.
[9] Szegedy C, Toshev A, Erhan D. Deep neural networks for object detection[C]. Advances in Neural Information Processing Systems. Lake Tahoe:Curran Associates Inc,2013:2553-2561.
[10] Ioffe S, Szegedy C. Batch normalization:Accelerating deep network training by reducing internal covariate shift[J]. Proc of the 32nd Int Conf on Machine Learning,2015, 37:448-456.
[11] He K, Zhang X, Ren S, et al. Delving deep into rectifiers:Surpassing human-level performance on imagenet classification[C]. Proc of the IEEE Int Conf on Computer Vision. Santiago:IEEE Computer Society Washington, 2015:1026-1034.
[12] Masci J, Meier U, Cirean D, et al. Stacked convolutional auto-encoders for hierarchical feature extraction[J]. Artificial Neural Networks and Machine Learning-ICANN, 2011, DOI:10.1007/978-3-642-21735-7_7.
[13] Cummins M, Newman P. Appearance-only SLAM at large scale with FAB-MAP 2.0[J]. The Int J of Robotics Research, 2011, 30(9):1100-1123.
[14] Gálvez-López D, Tardos J D. Bags of binary words for fast place recognition in image sequences[J]. IEEE Trans on Robotics, 2012, 28(5):1188-1197.
[15] Mei C, Sibley G, Newman P. Closing loops without places[C]. IEEE/RSJ Int Conf on IEEE Intelligent Robots and Systems(IROS). Taipei:IEEE Computer Society Washington, 2010:3738-3744.
[16] Murphy L, Sibley G. Incremental unsupervised topological place discovery[C]. IEEE Int Conf on Robotics and Automation(ICRA). Hong Kong:IEEE,2014:1312-1318.
[17] Kejriwal N, Kumar S, Shibata T. High performance loop closure detection using bag of word pairs[J]. Robotics and Autonomous Systems, 2016, 77:55-65.
[18] Khan S, Wollherr D. IBuILD:Incremental bag of binary words for appearance based loop closure detection[C].IEEE Int Conf on Robotics and Automation(ICRA).Seattle:IEEE, 2015:5441-5447.
[19]李博,杨丹,邓林.移动机器人闭环检测的视觉字典树金字塔TF-IDF得分匹配方法[J].自动化学报, 2011,37(6):665-673.(Li B, Yang D, Deng L. Visual vocabulary tree with pyramid TF-IDF scoring match scheme for loop closure detection[J]. Acta Automatica Sinica, 2011, 37(6):665-673.)
[20] Sünderhauf N, Shirazi S, Dayoub F, et al. On the performance of convnet features for place recognition[C].IEEE/RSJ Int Conf on Intelligent Robots and Systems(IROS). Hamburg:IEEE, 2015:4297-4304.
[21] Xu Yan, Tao Mo, Qiwei Feng, et al. Deep learning of feature representation with multiple instance learning for medical image analysis[C]. Int Conf on Acoustics, Speech and Signal Processing(ICASSP). Florence:IEEE, 2014:1626-1630.
[22] Gao X, Zhang T. Unsupervised learning to detect loops using deep neural networks for visual SLAM system[J].Autonomous Robots, 2017, 41(1):1-18.
[23] Xia Y, Li J, Qi L, et al. Loop closure detection for visual SLAM using PCANet features[C]. IEEE Int Joint Conf on Neural Networks. Vancouver:IEEE, 2016:2274-2281.
[24]罗杨宇,刘宏林.基于光束平差法的双目视觉里程计研究[J].控制与决策, 2016, 31(11):1936-1944.(Luo Y Y, Liu H L. Research on binocular vision odometer based on bundle adjustment method[J]. Control and Decision, 2016, 31(11):1936-1944.)
[25]季秀才,郑志强,张辉. SLAM问题中机器人定位误差分析与控制[J].自动化学报, 2008, 34(3):323-330.(Ji X C, Zheng Z Q, Zhang H. Analysis and control of robot position error in SLAM.[J]. Acta Automatica Sinica, 2008, 34(3):323-330.)
[26] Cummins M, Newman P. FAB-MAP:Probabilistic localization and mapping in the space of appearance[J].The Int J of Robotics Research, 2008, 27(6):647-665.
[27] Gao X, Zhang T. Loop closure detection for visual slam systems using deep neural networks[C]. The 34th Chinese Control Conf(CCC). Hangzhou:IEEE, 2015:5851-5856.
[28] Huang F J, Boureau Y L, LeCun Y. Unsupervised learning of invariant feature hierarchies with applications to object recognition[C]. IEEE Conf on Computer Vision and Pattern Recognition(CVPR). Minneapolis:IEEE, 2007:1-8.
[29] Ng A, Ngiam J, Foo C Y, et al. UFLDL tutorial[EB/OL].http://deeplearning.stanford.edu/wiki/index.php/UFLDL_Tutorial.
[2] Williams B, Klein G, Reid I. Automatic relocalization and loop closing for real-time monocular SLAM[J]. IEEE Trans on Pattern Analysis and Machine Intelligence,2011, 33(9):1699-1712.
[3] Kawewong A, Tongprasit N, Hasegawa O. A speeded-up online incremental vision-based loop-closure detection for long-term SLAM[J]. Advanced Robotics, 2013,27(17):1325-1336.
[4]赵洋,刘国良,田国会,等.基于深度学习的视觉SLAM综述[J].机器人, 2017, 39(6):889-896.(Zhao Y, Liu G L, Tian G H, et al. A survey of visual SLAM based on deep learning. robot[J]. Robot, 2017,39(6):889-896.)
[5] Ng P C, Henikoff S. SIFT:Predicting amino acid changes that affect protein function[J]. Nucleic Acids Research,2003, 31(13):3812-3814.
[6] Bay H, Ess A, Tuytelaars T, et al. Speeded-up robust features(SURF)[J]. Computer Vision and Image Understanding, 2008, 110(3):346-359.
[7] Shekhar R, Jawahar C V. Word image retrieval using bag of visual words[C]. IEEE 10th IAPR Int Workshop on Document Analysis Systems(DAS). Gold Coast:IEEE,2012:297-301.
[8] Krizhevsky A, Sutskever I, Hinton G E. Imagenet classification with deep convolutional neural networks[C]. Advances in neural information processing systems. Lake Tahoe:Curran Associates Inc, 2012:1097-1105.
[9] Szegedy C, Toshev A, Erhan D. Deep neural networks for object detection[C]. Advances in Neural Information Processing Systems. Lake Tahoe:Curran Associates Inc,2013:2553-2561.
[10] Ioffe S, Szegedy C. Batch normalization:Accelerating deep network training by reducing internal covariate shift[J]. Proc of the 32nd Int Conf on Machine Learning,2015, 37:448-456.
[11] He K, Zhang X, Ren S, et al. Delving deep into rectifiers:Surpassing human-level performance on imagenet classification[C]. Proc of the IEEE Int Conf on Computer Vision. Santiago:IEEE Computer Society Washington, 2015:1026-1034.
[12] Masci J, Meier U, Cirean D, et al. Stacked convolutional auto-encoders for hierarchical feature extraction[J]. Artificial Neural Networks and Machine Learning-ICANN, 2011, DOI:10.1007/978-3-642-21735-7_7.
[13] Cummins M, Newman P. Appearance-only SLAM at large scale with FAB-MAP 2.0[J]. The Int J of Robotics Research, 2011, 30(9):1100-1123.
[14] Gálvez-López D, Tardos J D. Bags of binary words for fast place recognition in image sequences[J]. IEEE Trans on Robotics, 2012, 28(5):1188-1197.
[15] Mei C, Sibley G, Newman P. Closing loops without places[C]. IEEE/RSJ Int Conf on IEEE Intelligent Robots and Systems(IROS). Taipei:IEEE Computer Society Washington, 2010:3738-3744.
[16] Murphy L, Sibley G. Incremental unsupervised topological place discovery[C]. IEEE Int Conf on Robotics and Automation(ICRA). Hong Kong:IEEE,2014:1312-1318.
[17] Kejriwal N, Kumar S, Shibata T. High performance loop closure detection using bag of word pairs[J]. Robotics and Autonomous Systems, 2016, 77:55-65.
[18] Khan S, Wollherr D. IBuILD:Incremental bag of binary words for appearance based loop closure detection[C].IEEE Int Conf on Robotics and Automation(ICRA).Seattle:IEEE, 2015:5441-5447.
[19]李博,杨丹,邓林.移动机器人闭环检测的视觉字典树金字塔TF-IDF得分匹配方法[J].自动化学报, 2011,37(6):665-673.(Li B, Yang D, Deng L. Visual vocabulary tree with pyramid TF-IDF scoring match scheme for loop closure detection[J]. Acta Automatica Sinica, 2011, 37(6):665-673.)
[20] Sünderhauf N, Shirazi S, Dayoub F, et al. On the performance of convnet features for place recognition[C].IEEE/RSJ Int Conf on Intelligent Robots and Systems(IROS). Hamburg:IEEE, 2015:4297-4304.
[21] Xu Yan, Tao Mo, Qiwei Feng, et al. Deep learning of feature representation with multiple instance learning for medical image analysis[C]. Int Conf on Acoustics, Speech and Signal Processing(ICASSP). Florence:IEEE, 2014:1626-1630.
[22] Gao X, Zhang T. Unsupervised learning to detect loops using deep neural networks for visual SLAM system[J].Autonomous Robots, 2017, 41(1):1-18.
[23] Xia Y, Li J, Qi L, et al. Loop closure detection for visual SLAM using PCANet features[C]. IEEE Int Joint Conf on Neural Networks. Vancouver:IEEE, 2016:2274-2281.
[24]罗杨宇,刘宏林.基于光束平差法的双目视觉里程计研究[J].控制与决策, 2016, 31(11):1936-1944.(Luo Y Y, Liu H L. Research on binocular vision odometer based on bundle adjustment method[J]. Control and Decision, 2016, 31(11):1936-1944.)
[25]季秀才,郑志强,张辉. SLAM问题中机器人定位误差分析与控制[J].自动化学报, 2008, 34(3):323-330.(Ji X C, Zheng Z Q, Zhang H. Analysis and control of robot position error in SLAM.[J]. Acta Automatica Sinica, 2008, 34(3):323-330.)
[26] Cummins M, Newman P. FAB-MAP:Probabilistic localization and mapping in the space of appearance[J].The Int J of Robotics Research, 2008, 27(6):647-665.
[27] Gao X, Zhang T. Loop closure detection for visual slam systems using deep neural networks[C]. The 34th Chinese Control Conf(CCC). Hangzhou:IEEE, 2015:5851-5856.
[28] Huang F J, Boureau Y L, LeCun Y. Unsupervised learning of invariant feature hierarchies with applications to object recognition[C]. IEEE Conf on Computer Vision and Pattern Recognition(CVPR). Minneapolis:IEEE, 2007:1-8.
[29] Ng A, Ngiam J, Foo C Y, et al. UFLDL tutorial[EB/OL].http://deeplearning.stanford.edu/wiki/index.php/UFLDL_Tutorial.