| 基于LVQ神经网络的目标识别技术 |
| |
| 引用本文： | 黄勇,刘燕. 基于LVQ神经网络的目标识别技术[J]. 光学与光电技术, 2006, 4(1): 58-61 |
| |
| 作者姓名： | 黄勇 刘燕 |
| |
| 作者单位： | 华中光电技术研究所,武汉,430073;华中光电技术研究所,武汉,430073 |
| |
| 摘 要： | 通过对目标特征的分析比较,选取不变矩作为识别特征.为了达到很好的识别效果,对不变矩做了优化处理.采用LVQ神经网络技术建立了识别模型,提高了识别速度.运用有限的样本对目标识别技术进行了测试,结果表明采用此技术后的识别成功率较高,平均约为98%,识别速度快,每幅二值目标图像的识别时间约为16 ms.
|
| 关 键 词： | 目标识别 目标特征 中心矩 不变矩 神经网络 |
| 文章编号： | 1672-3392（2006）01-0058-04 |
| 收稿时间： | 2005-08-20 |
| 修稿时间： | 2005-10-09 |
Target Recognition Based on LVQ Neural Network |
| |
| HUANG Yong,LIU Yan. Target Recognition Based on LVQ Neural Network[J]. optics&optoelectronic technology, 2006, 4(1): 58-61 |
| |
| Authors: | HUANG Yong LIU Yan |
| |
| Abstract: | By the analysis and comparison of target characteristics, invariant moments are chosen as recognition characteristics. In order to get good recognition, invariant moments are transformed and optimized. The limited samples are used to test the target recognition technique. The test result shows that the recognition probability is as high as 98% and the recognition time is as short as 16 ms. |
| |
| Keywords: | target recognition target characteristics center moment invariant moment neural network |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
