基于小波的图像分割方法.doc

基于小波的图像分割方法 【摘 要】： 近年来随着多媒体技术的发展，图像技术也得到了极大的重视和发展，从而这就促成了图像技术学科的发展。在图像技术中，图像分割是得到图像分析的关键步骤而图像分析的目标是要靠图像分割技术提取出来的；图像的分割、特征的提取和参数的测量，将原始图像转化为更为抽象和紧凑的形式，简化了问题，同时提取到图的图像压缩与编码技术中，图像分割也是一个重要的步骤。 传统的图像分割方法主要是基于图像的灰度特征的。分割算法可分为利用区域间灰度不连续性的基于边缘的算法和利用区域内灰度相似性的基于阈值的算法。 本文首先介绍了基于小波的图像分割有关理论和方法。然后使用该方法对图像的灰度直方图进行小波多尺度变换，并从较大的尺度系数到较小的尺度系数逐步定位出灰度阈值。通过实验可知该方法具有良好的抗噪声性能。 【关键词】： 图像处理，波变换，尺度分析，图像分割 Abstract In recent years along with multimedia technologies development, the image technology also obtained the enormous value and the development, thus this has facilitated the image technology discipline development. In the image technology, the image division obtains the image analysis committed step, but image analysiss goal is must depend on the image division technology to withdraw; The image division, the characteristic extraction and the parameter survey, transforms the primitive image as more abstract and a compact form, simplified the question, simultaneously withdraws in the image compression and the coding technique, the image division is also one important step. The traditional image division method is mainly based on the image gradation characteristic. The division algorithm may divide into uses the regional gradation discontinous and uses in the region based on the edge algorithm the gradation similar based on the threshold value algorithm. This article first introduced based on the wavelet image division related theory and the method. Then uses this method to carry on the young Pood scaling transform for the image gradation histogram, and locates the gradation threshold value gradually from the great scale coefficient to the small scale coefficient. May know this method through the experiment to have the good anti-noise performance. Key words：Imagery processing， Wavelet transformation;，Multi-criterion analysis， Image division 1．引 言 1 1.1小波分析发展史 2 1.2小波在图像分割中的应用 4 2．小波变换 5 2.1小波变换的定义 6 2.2小波变换的性质 8 2.3小波多分辨分析 9 2.4小波分析的算法 10 2.5多进制小波变换概述 12 3．边缘检测方法 13 3.1基于小波分析的多尺度边