一种机器人室内场景图像识别方法

J4 ›› 2017, Vol. 34 ›› Issue (3): 293-304.

一种机器人室内场景图像识别方法

刘畅

1 沈阳职业技术学院，辽宁沈阳 110045； 2东北大学信息科学与工程学院，辽宁沈阳 110004

收稿日期:2015-11-02 修回日期:2016-03-31 出版日期:2017-05-28 发布日期:2017-05-22
通讯作者: 刘畅（1977-），女，辽宁昌图人，主要研究方向为双目识别、机器视觉级计算机软件开发。
基金资助:
Supported by National Natural Science Foundation of China(国家自然科学基金，61273078)

A quantization Robot indoor scene image recognition method

1 Shengyang Polytechnic College, Shenyang 110045, China; 2 School of Information Science and Engineering，Northeastern University，Shenyang 110004，China

Received:2015-11-02 Revised:2016-03-31 Published:2017-05-28 Online:2017-05-22

摘要/Abstract

摘要： 为了对现有机器人的物体识别进行优化和改进，提出了一种新的权重计算方法进行室内场景图像识别。该方法通过对输入场景的转换获取无向带权图，在表面法方向的基础上，使用表面凹凸度这一指标来进行表面凹凸度综合判定，取代了传统的布尔判定，大大提升了抗噪性能，避免出现错误传递放大的情况。基于快速图像分割算法及时识别未知物体。实验结果表明：提出方法的鲁棒性及抗噪能力均较强，优于单纯基于法方向的方法。与基于深度学习与推测的同类方法相比，所提方法性能更好，更适用于实际识别。

关键词: 权重；凹凸度；快速图像分割算法；布尔判断；图像识别

Abstract: In order to optimize and improve object recognition of the existing robots, a new weight calculation method for interior scene images identification is proposed. The undirected weighted graph is obtained by converting the input scene. Based on surface normal direction, comprehensive determination of surface roughness is carried out using the concave and convex degree index instead of the traditional Boolean decision, which greatly enhances the anti noise performance and avoids the error propagation amplification. The unknown objects are identified in time based on fast image segmentation algorithm. Experiment results show that the robustness and anti noise ability of the method proposed are both strong, and the method proposed is better that methods based on normal direction only. Compared with other methods based on deep leaning and conjecture, the method proposed has better performance, and it’s more applicable to the practical identification.

Key words: weight; concave and convex degree; fast image partitioning algorithm; Boolean decision; image identification

刘畅. 一种机器人室内场景图像识别方法[J]. J4, 2017, 34(3): 293-304.

参考文献

[1] Vasudevan S, G?chter S, Nguyen V, et al. Cognitive maps for mobile robots—an object based approach[J]. Robotics and Autonomous Systems, 2007, 55(5):359–371.
[2] Rusu R B, Cousins S. 3d is here: Point cloud library (pcl)[A]. In: Robotics and Automation (ICRA),2011 IEEE International Conference on[C], 2011. 1–4.
[3] Schnabel R, Wahl R, Klein R. Efficient RANSAC for Point-Cloud Shape Detection[A]. In: Computergraphics forum[C], 2007. 26:214–226.
[4] Klasing K, Wollherr D, Buss M. A clustering method for efficient segmentation of 3D laser data.[A].In: ICRA[C], 2008. 4043–4048.
[5] Triebel R, Shin J, Siegwart R, et al. Segmentation and unsupervised part-based discovery of repetitive objects[A]. In: Robotics: Science and Systems[C], 2010. 2.
[6] Moosmann F, Pink O, Stiller C. Segmentation of 3D lidar data in non-flat urban environments using a local convexity criterion[A]. In: Intelligent Vehicles Symposium, 2009 IEEE[C], 2009. 215–220.
[7]SteinSC,W?rg?tterF,SchoelerM,etal.Convexitybasedobjectpartitioningforrobotapplications[A].In:Robotics and Automation (ICRA), 2014 IEEE International Conference on[C], 2014. 3213–3220.
[8] He K, Sun J, Tang X. Guided image filtering[M]//. In: Computer Vision–ECCV 2010[C].[S.l.]: Springer, 2010. 1–14.
[9] Trevor A, Gedikli S, Rusu R, et al. Efficient organized point cloud segmentation with connected components[J]. Semantic Perception Mapping and Exploration (SPME), 2013.
[10] Poppinga J, Vaskevicius N, Birk A, et al. Fast plane detection and polygonalization in noisy 3D range images[A]. In: Intelligent Robots and Systems, 2008. IROS 2008. IEEE/RSJ International Conference on[C], 2008. 3378–3383.
[11] Holz D, Behnke S. Fast range image segmentation and smoothing using approximate surface recon-struction and region growing[M]//. In: Intelligent Autonomous Systems 12[C].[S.l.]: Springer, 2013.61–73.
[12] Rabbani T, van den Heuvel F, Vosselmann G. Segmentation of point clouds using smoothness con-straint[J]. International Archives of Photogrammetry, Remote Sensing and Spatial Information Sci-ences, 2006, 36(5):248–253.
[13] Uckermann A, Haschke R, Ritter H. Real-time 3D segmentation of cluttered scenes for robot grasp-ing[A]. In: Humanoid Robots (Humanoids), 2012 12th IEEE-RAS International Conference on[C],2012. 198–203.
[14] Holz D, Holzer S, Rusu R B, et al. Real-time plane segmentation using RGB-D cameras[M]//. In:RoboCup 2011: Robot Soccer World Cup XV[C].[S.l.]: Springer, 2012. 306–317.
[15] Holzer S, Rusu R B, Dixon M, et al. Adaptive neighborhood selection for real-time surface normal estimation from organized point cloud data using integral images[A]. In: Intelligent Robots and Systems (IROS), 2012 IEEE/RSJ International Conference on[C], 2012. 2684–2689.
[16] Karpathy A, Miller S, Fei-Fei L. Object discovery in 3D scenes via shape analysis[A]. In: Robotics and Automation (ICRA), 2013 IEEE International Conference on[C], 2013. 2088–2095.
[17] Felzenszwalb P F, Huttenlocher D P. Efficient graph-based image segmentation[J]. International Journal of Computer Vision, 2004, 59(2):167–181.
[18] Richtsfeld A. The Object Segmentation Database (OSD)[G]. http://www.acin.tuwien.ac.at/?id=289, 2012.
[19] Richtsfeld A, Morwald T, Prank I J, etal. Segmentation of unknown object in indoor environment[A].In:Intelligent Robots and System (IROS), 2012 IEEE/RSJ International Conference on[C], 2012.4791–4796.