摘要
在利用数字射线成像系统进行特征检测过程中,由于系统和工件本身特点的原因,在获取的射线图像中,某些需要识别的特征信息常常被大量的低频背景信息掩盖,使特征的识别增加了难度。针对上述问题,研究了一种基于经验模式分解(Empirical Mode Decomposition,简称EMD)的图像特征增强技术。研究表明,与传统分析方法相比,EMD方法能够更准确地剔除图像中存在的非需特征信息,增强需要识别的特征信息,提高特征信息的识别率。
In the process of feature recognition that using the digital radiographic system, because the system and the workpiece have themselves characteristics and within the ray image captured , some feature information that need to be recognized are often covered up by large number of low-frequency background information, which increases the target-recognizing difficulty. In view of the above questions, an empirical mode decomposition (EMD) based on approaching for image feature enhancing is investigated in this article. Comparing with the traditional analysis methods ,the results show that the approach can eliminate needless features information of images more precisely, enhance image feature information which need to recognized and ncrease the characteristic-recognition rate.
出处
《山西电子技术》
2009年第4期34-36,共3页
Shanxi Electronic Technology
关键词
经验模式分解
特征增强
特征识别
empirical mode decomposition
feature enhancing
feature recognition