摘要
针对传统的机器人视觉目标图像处理中将边缘信息与噪声均作为噪声处理,从而使机器人无法精确完成图像识别的问题,提出采用一种改进中值滤波与改进邻域均值滤波相结合的滤波算法来移除噪声。首先采用改进的中值滤波算法,用滑动大窗口分块处理来加快运行速度,并加入判断条件,区别边缘与噪声信息,避免边缘信息丢失,并去除噪声点;然后采用改进邻域均值滤波方法对图像边缘信息进行处理,从而保护好图像的边缘信息。仿真分析结果表明,该算法能较好地滤除噪声,明显优于经典的中值滤波与邻域均值滤波算法。
In traditional robot vision target image processing,the edge information and noise are both used as noise,so the robot cannot recognize the image accurately.To deal with this problem,a filtering algorithm of the combination of an improved median filter and an improved neighborhood averaging filter is adopted to remove the image noise in the robot image recognition system.First,the improved median filter by a sliding big window block method is used to accelerate the speed.Then,in order to distinguish noises from edge informations,the conditions are added to avoid the edge information to be lost and to remove the noise points.Finally,the improved neighborhood averaging filter is used to process and protect the edge imformation.Simulation results show that the algorithm can remove the noises better,and is obviously better than the classical median filter algorithm and the neighborhood averaging filter algorithm.
出处
《机械科学与技术》
CSCD
北大核心
2011年第5期823-826,832,共5页
Mechanical Science and Technology for Aerospace Engineering
基金
湖南省教育厅一般项目(08C868)资助
关键词
机器人视觉目标
图像处理
改进中值滤波
改进邻域均值滤波
robot vision target
image processing
improved median filter
improved neighborhood averaging filter