J4 ›› 2006, Vol. 28 ›› Issue (11): 136-139.
• 论文 • 上一篇 下一篇
李科[1,2] 王润生[2] 王程[2]
出版日期:
发布日期:
Online:
Published:
摘要：
本文提出了一种飞机图像目标识别的方法。首先，我们建立了由多种型号飞机在大小、旋转角度改变或进行简单仿射变换等情况下的飞机样本图像库；其次，将样本库分为训 练样本库和测试样本库，提取了训练样本库中飞机图像的不变矩、仿射不变矩、机长翼展比和紧凑度等特征量，对这些特征量进行了分析，建立了飞机目标的表述模型；最后后，分别用最小距离、BP神经网络和树分类器进行了分类实验。实验表明，树分类器方法效果较好。
关键词: 目标识别 最小距离 神经网络 树分类器
Abstract:
In this paper, a method for the recognition of airplane images is proposed. First, we establish a swatch database including a training swatch database and a testing swatch database with many types of airplanes＇ planforms under the condition of different scales, different view angles and simple affine e transformations. And then, we extract the Hu invariant moment, the affine invariant moment, the ratio of length to width and the compactness degree as the characteristics of airplane targets, and analyze these characteristics. Finally, we experiment with the least-distance classifier, the neural netwo rk classifier and the tree classifier. As a result, the tree classifier is better than the other two.
Key words: （target recognition, least-distance, neural network, tree classifier）
李科[1,2] 王润生[2] 王程[2]. 一种用于飞机型号识别的树分类器方法[J]. J4, 2006, 28(11): 136-139.
0 / / 推荐
导出引用管理器 EndNote|Ris|BibTeX
链接本文: http://joces.nudt.edu.cn/CN/
http://joces.nudt.edu.cn/CN/Y2006/V28/I11/136
