【摘要】 在基于电动轮椅的机械臂目标抓取研究背景下,通过图像分割实现对目标物的定位.只有目标物与背景的准确分离,才能使机械臂通过轮椅移动平台准确地移动到其工作范围内实现抓取工作.结合研究中目标物受光线因素影响而导致图像分割不准确现象,本文提出了一种基于区域生长的PCNN自动目标分割算法.经实验证明,该方法可以有效避免传统PCNN算法因目标物受光线因素影响而无法对目标物准确分割的问题,且该方法通过目标物分割而获得定位坐标的准确率达到98%,仅对感兴趣区域进行分割,避免对无关信息的处理,运行效率平均达1.61s,具有一定实时性.更多还原

【Abstract】 In the research of robotic arm target capture with electric wheelchair,target location can be realized through image segmentation.Accurate separation of target and background is the premise of accurate target positioning,which can make the arm move to the workspace to realize target capture accurately by the chair.To solve the problem that light factor can make the image segmentation inaccurate,a localization algorithm is proposed based on growth reign and PCNN.The experiment results show that this algorithm can effectively realize what the traditional PCNN algorithm can not do,and the localization accuracy reaches 98%.As the new algorithm can help avoid irrelevant information processing,it can increase the dealing speed to about 1.61 s.更多还原