| 复杂背景中的目标辨识方法研究 |
| |
| 引用本文： | 张天序,李小文,彭嘉雄,邓小先.复杂背景中的目标辨识方法研究[J].华中科技大学学报(自然科学版),1994(5). |
| |
| 作者姓名： | 张天序 李小文 彭嘉雄 邓小先 |
| |
| 作者单位： | 华中理工大学图像识别与人工智能研究所 |
| |
| 摘 要： | 研究了二维熵准则极值的递推算法，以获取图像分割的最优门限，基于知识和多特征复合的思想，提出了一种具有知觉组织特点的目标辨识方法，给出了识别复杂场景中特定目标的实验结果。
|
| 关 键 词： | 多特征提取，知觉组织，知识，机器视觉，目标辨识 |
Identification of Objects in a Complex Background:a Case Study |
| |
| Zhang TianxuInst of Pattern Recognition & AI,H.U.S.T.,Wuhan ,China.,Li Xiaowen,Peng Jiaxiong,Deng Xiaoxian.Identification of Objects in a Complex Background:a Case Study[J].JOURNAL OF HUAZHONG UNIVERSITY OF SCIENCE AND TECHNOLOGY.NATURE SCIENCE,1994(5). |
| |
| Authors: | Zhang TianxuInst of Pattern Recognition & AI HUST Wuhan China Li Xiaowen Peng Jiaxiong Deng Xiaoxian |
| |
| Institution: | Zhang TianxuInst of Pattern Recognition & AI,H.U.S.T.,Wuhan 430074,China.,Li Xiaowen,Peng Jiaxiong,Deng Xiaoxian |
| |
| Abstract: | The reason why the recognition of an object in a complex background is very difficult inmachine vision application is that the object can hardly be extracted intact from its back-ground.Afast recursive algorithm for calculating the extreme value of the two-dimensionalentropy criterion to determine the optimum threshold of image segmentation is studied.Based on the oUect knowledge and multifeature fusion,a new method for object identifica-tion that features the perceptual organization is proposed.The intact extraction and definiteidentification of the object are achieved almost at the same time. Experimental results of i-dentifying some specific objects in a complicated scene are given. |
| |
| Keywords: | multifeature extraction perceptual organization knowledge machine vision object identification |
| 本文献已被 CNKI 等数据库收录！ |
|
