摘要
为了满足移动机器人准确快速的目标识别要求,在利用Gabor滤波器识别目标前,给出一种改进的区域生长算法.该方法可自动搜索和确定生长点,避免了区域生长法有时会出现的"死循环"现象.该方法计算过程简单,虽然误识率仍较高,但拒识率低,可用于目标识别的预处理,提高识别的实时性.
To improve accuracy and speed of target recognition of mobile robot,a method of improved region growing algorithm used before Gabor filter was given.The method can search and locate growing points automatically,which avoid ＂endless loop＂ sometimes by region growing.Although the method has higher misclassification rate,it has simple calculation process and good rejection performance.The method can improve real-time performance of recognition,when it is used on pretreatment of target recognition.
出处
《微电子学与计算机》
CSCD
北大核心
2010年第10期124-127,共4页
Microelectronics & Computer
基金
国家"八六三"计划项目(2009AA04Z161)
江苏省"六大人才高峰"项目(07-E-024)
关键词
移动机器人
区域生长
区域合并
目标识别
mobile robot
region growing
region merging
target recognition
