文档介绍：基于小波的图像分割方法研究
摘要
在过去的几十年里,计算机系统的处理速度,存储能力以及计算机网络都飞速发展着,这些惊人的发展都为数字图像处理以及网络多媒体都提供了有力的技术保障,并且极大的拓展了数字图像处理及网络多媒体的研究和应用。数字图像处理已经成为一个令人瞩目的学科项目,其研究领域和研究课题都数不胜数。数字图像处理理论和技术大致可概括为:图像数字化、图像恢复、图像增强、图像压缩、图像重建、图像分割、特征提取等。这篇文章的研究重点在于基于小波的图像分割技术。首先图像分割的关键步骤是由图像处理到图像分析,在图像工程中占据重要的位置。然而传统方法中还存在着很多不足,经过本文的分析和研究我们得出以下结论:传统的直方图阈值法图像分割技术的基本思想是图像由不同特征区域构成,它的直方图呈现出多个峰值,每个峰值相应地对应着一个区域,以谷值点为阈值划分相邻峰值。如果直方图有明显的谷值点,则能比较容易选择出阈值,然而在实际应用中,由于噪声的存在,很难准确地选择阈值,阈值设定容易受噪声的影响。基于小波变换的阈值法图像分割技术则能够有效地避免噪声的影响。图像分割技术的主导思想是首先由二进制小波变换将图像的直方图分解为不同层次的小波系数,随后依据给定的分割准则和小波系数选择阈值门限,最后利用阈值标出图像分割的区域。与以往方的图像分割方式相比有明显较大的优势。
关键词
小波;哈尔小波;图像分割;边缘检测;matlab仿真
Abstract
In the past years, the processing speed of puter,storage power work are all developing with an amazing speed . The great development provide a strong technical support for the digital image processing work multimedia ,also expand the research and processing is ing a booming science projects,it has a wide rang of reserching area and processing theory and technology include:image digitization 、image enhancement 、image pression 、image segmentation 、 feature extraction .The focus of this article is mage segmentation .First,image segmentation is the key steps from image processing to image analysis , and in the image works occupy an important There are a lot ings than traditional methods,After this analysis we found the following findings:The traditional histogram threshold method The basic idea of the image segmentation is the image constitute the different feature regions, which show multi-peak histogram, each peak corresponding to a region to the threshold point valley by neighboring the histogram valley obvious point, is relatively easy to select the threshold value, but in practice, due to the presence of noise, is often difficult to effectively select the threshold, the threshold value is set to sensitive noise. Wavelet-based image segmentationthreshold method is able to effectively avoid the noise. The