基于梯度图像和模板匹配的单目视觉障碍物检测方法
|
摘要
障碍物检测是自动驾驶、机器人自主导航的核心问题之一。为了检测障碍物,提出一种基于聚类和目标匹配的单目视觉方法。在机器人使用场景下,道路的纹理、色彩具有相似或者一致性,在这样的假设下,对可能的道路点和非道路点进行多次取样,通过聚类的方法找出道路点的特征,然后进行匹配,即可标记出全部的道路点,未被标记的点即为障碍物,从而将障碍物检测出来。匹配时考虑图像梯度,将平均匹配准确度提高到90%以上。
Obstacle detection is one of the core issues of autonomous driving and robot autonomous navigation.A monocular vision method is proposes based on clustering and template matching to detect obstacles.In the robot applying scenario,the texture and color of the road are similar or consistent.Under this assumption,the possible road points and non-road points are sampled multiple times,and the characteristics of the road points are found through the clustering method.Matching will mark all the road points,and unmarked points will be obstacles obviously.Considering the image gradient when matching,the average matching accuracy has been improved to over 90%.
Obstacle detection is one of the core issues of autonomous driving and robot autonomous navigation.A monocular vision method is proposes based on clustering and template matching to detect obstacles.In the robot applying scenario,the texture and color of the road are similar or consistent.Under this assumption,the possible road points and non-road points are sampled multiple times,and the characteristics of the road points are found through the clustering method.Matching will mark all the road points,and unmarked points will be obstacles obviously.Considering the image gradient when matching,the average matching accuracy has been improved to over 90%.
引文
[1] 王葵,翟荣刚,王道斌,等.基于激光测距雷达和车载GPS的动态障碍物检测[J].工业仪表与自动化装置,2013,(2):10-13.
[2] 辛煜,梁华为,梅涛,等.基于激光传感器的无人驾驶汽车动态障碍物检测及表示方法[J].机器人,2014,36(6):654-661.
[3] 王新竹,李骏,李红建,等.基于三维激光雷达和深度图像的自动驾驶汽车障碍物检测方法[J].吉林大学学报(工学版),2016,46(2)360-365.
[4] GAGEIK N,BENZ P,MONTENEGRO S.Obstacle Detection and Collision Avoidance for a UAV With Complementary Low-Cost Sensors[J].IEEE Access,2015,3:599-609.
[5] DANIEL MAIER,CYRILL STACHNISS,MAREN BENNEWITZ.VISION-BASED HUMANOID NAVIGATION USING SELF-SUPERVISED OBSTACLE DETECTION[J].International Journal of Humanoid Robotics,2013,10(02):90-216.
[6] 李丽军.基于机器人双目立体视觉的三维重建[D].太原理工大学,2012.
[7] 何祖恩,吴海彬,李文锦.基于双目视觉的移动机器人障碍物检测研究[J].机械设计与制造,2009,(12):171-173.
[8] POLLINI L,GRECO F,MATI R,et al.Stereo Vision Obstacle Detection Based on Scale Invariant Feature Transform Algorithm[C]// AIAA Guidance,Navigation and Control Conference and Exhibit,2013.
[9] SUN H,ZOU H,ZHOU S,et al.Surrounding Moving Obstacle Detection for Autonomous Driving Using Stereo Vision[J].International Journal of Advanced Robotic Systems,2013,10(10):1-8.
[10] KRISTAN M,SULIC K V,KOVACIC S,et al.Fast Image-Based Obstacle Detection From Unmanned Surface Vehicles.[J].IEEE Trans Cybern,2015,46(3):641-654.
[11] 张磊,史金飞,罗翔.运动目标检测的差分图像法分析研究[J].工业仪表与自动化装置,2007,2007(3):7-11.
[12] 吴振杰,毛晓波.一种改进的帧间差光流场算法[J].计算机工程与应用,2013,49(18):200-203.
[13] ZHENG Lai-fang,SUN Wei,OUYang Ming-hua et al.Obstacle detection method for duct cleaning robot combining optical flow and artificial potential field[J].Computer Engineering and Applications,2016,52(9):243-247.
[14] 袁周米琪,周坚华.自适应确定K-means算法的聚类数:以遥感图像聚类为例[J].华东师范大学学报(自然科学版),2014,(6):73-80.
[15] 王哲峰.基于提升小波变换的图像匹配算法研究[D].吉林大学,2007.
[16] PRAKASH O,SRIVASTAVA R,KHARE A.Biorthogonal wavelet transform based image fusion using absolute maximum fusion rule[C]// Information & Communication Technologies.IEEE,2013:577-582.
[17] 程红,陈文剑,孙文邦.一种改进的快速归一化积相关图像匹配算法[J].光电工程,2013,40(1):118-125.
[18] 韩冰,王永明.基于一种快速归一化积相关算法图像匹配研究[J].兵工学报,2010,31(2):160-165.
[19] FURNARI A,FARINELLA G M,BRUNA A R,et al.Generalized Sobel Filters for gradient estimation of distorted images[C]// IEEE International Conference on Image Processing.IEEE,2015:3250-3254.
[20] 孔岩,管庶安,江峰,等.基于二阶梯度的图像增强算法研究[J].计算机与数字工程,2011,39(4):137-141.
[2] 辛煜,梁华为,梅涛,等.基于激光传感器的无人驾驶汽车动态障碍物检测及表示方法[J].机器人,2014,36(6):654-661.
[3] 王新竹,李骏,李红建,等.基于三维激光雷达和深度图像的自动驾驶汽车障碍物检测方法[J].吉林大学学报(工学版),2016,46(2)360-365.
[4] GAGEIK N,BENZ P,MONTENEGRO S.Obstacle Detection and Collision Avoidance for a UAV With Complementary Low-Cost Sensors[J].IEEE Access,2015,3:599-609.
[5] DANIEL MAIER,CYRILL STACHNISS,MAREN BENNEWITZ.VISION-BASED HUMANOID NAVIGATION USING SELF-SUPERVISED OBSTACLE DETECTION[J].International Journal of Humanoid Robotics,2013,10(02):90-216.
[6] 李丽军.基于机器人双目立体视觉的三维重建[D].太原理工大学,2012.
[7] 何祖恩,吴海彬,李文锦.基于双目视觉的移动机器人障碍物检测研究[J].机械设计与制造,2009,(12):171-173.
[8] POLLINI L,GRECO F,MATI R,et al.Stereo Vision Obstacle Detection Based on Scale Invariant Feature Transform Algorithm[C]// AIAA Guidance,Navigation and Control Conference and Exhibit,2013.
[9] SUN H,ZOU H,ZHOU S,et al.Surrounding Moving Obstacle Detection for Autonomous Driving Using Stereo Vision[J].International Journal of Advanced Robotic Systems,2013,10(10):1-8.
[10] KRISTAN M,SULIC K V,KOVACIC S,et al.Fast Image-Based Obstacle Detection From Unmanned Surface Vehicles.[J].IEEE Trans Cybern,2015,46(3):641-654.
[11] 张磊,史金飞,罗翔.运动目标检测的差分图像法分析研究[J].工业仪表与自动化装置,2007,2007(3):7-11.
[12] 吴振杰,毛晓波.一种改进的帧间差光流场算法[J].计算机工程与应用,2013,49(18):200-203.
[13] ZHENG Lai-fang,SUN Wei,OUYang Ming-hua et al.Obstacle detection method for duct cleaning robot combining optical flow and artificial potential field[J].Computer Engineering and Applications,2016,52(9):243-247.
[14] 袁周米琪,周坚华.自适应确定K-means算法的聚类数:以遥感图像聚类为例[J].华东师范大学学报(自然科学版),2014,(6):73-80.
[15] 王哲峰.基于提升小波变换的图像匹配算法研究[D].吉林大学,2007.
[16] PRAKASH O,SRIVASTAVA R,KHARE A.Biorthogonal wavelet transform based image fusion using absolute maximum fusion rule[C]// Information & Communication Technologies.IEEE,2013:577-582.
[17] 程红,陈文剑,孙文邦.一种改进的快速归一化积相关图像匹配算法[J].光电工程,2013,40(1):118-125.
[18] 韩冰,王永明.基于一种快速归一化积相关算法图像匹配研究[J].兵工学报,2010,31(2):160-165.
[19] FURNARI A,FARINELLA G M,BRUNA A R,et al.Generalized Sobel Filters for gradient estimation of distorted images[C]// IEEE International Conference on Image Processing.IEEE,2015:3250-3254.
[20] 孔岩,管庶安,江峰,等.基于二阶梯度的图像增强算法研究[J].计算机与数字工程,2011,39(4):137-141.