基于SVM梅花种类识别系统的设计+源代码
毕业论文关键词：梅花种类识别；分割算法；特征提取；支持向量机（SVM）
The design and implementation of the type of plum blossom Recognition system based on SVM
Abstract: Nowadays,digital image processing,digital image analysis, and pattern recognition technology play an important part in computer science and technology. As one of the digital image processing technologies,the technology of automatic classification and recognition of plants has been widely used in daily life.Plum blossom was studied as research project in this paper.Image segmentation, feature extraction,classification, etc. The key work and contributions in this papers are as follow:firstly,Plant species identification technologies including texture segmentation and color segmentationbased on image analysis were analyzed and researched deeply. Double blanket algorithm were applied in texture segmentation.FCM Clustering Algorithm were applied in color segmentation.secondly,On feature extraction, the local feature extraction was researched combing with the result of texture feature, shape features, and color features.Classification was researched using a Support Vector Machine (SVM) based on thefeatures extracted from the feature extraction method.
Key words:Type recognition of plum blossom;Segmentation algorithm;Feature extraction ;Support vector machine (SVM)
摘要 1
关键词 1
Abstract 1
Key words 1
引言 1
1 绪论 1
1．1 研究背景与意义 1
1．2国内外发展及研究现状 2
1.3 研究内容与技术路线 2
1.3.1 研究内容 2
1.3.2 技术路线 2
2开发平台及应用技术 5
2.1开发平台介绍 5
2.2 SVM介绍 5
3梅花图像的分割提取算法 6
3.1纹理分割 6
3.1.1双毯子算法计算分形文数 6
3.1.2大津阈值分割 7
3.1.3填充，去噪，逻辑与 7
3.2颜色分割 7
3.2.1色彩空间的转换 7
3.2.2 FCM算法分割 7
3.2.3后续处理 8
3.3 梅花图像的特征提取 8
3.3.1颜色特征提取 8
3.3.1形状特征提取 9
3.3.2纹理特征提取 10
3.3.3局部区块特征提取 10
4系统设计与实现 12
4.1图像分割 12
4.1特征提取 12
4.4图像分类识别 13
5讨论和总结 17
6致谢 17
参考文献： 17
基于SVM的梅花种类识别系统的设计与实现
引言：从事园艺方面工作的专业技术人员可以依靠自己所学过的专业知识或者靠长期工作积累下来的经验来识别各种各样的植物的种类，但是植物种类有上万种，靠人工识别，准确率太低。而对于非专业的人员来说，区分花卉的种类就更加困难了。而计算机结合图像处理的模式识别技术解决人工识别效率低下的问题。由于梅花在花卉中比较特殊，在开花的时候背景中没有叶子的干扰，有利于梅花图像的分割提取。本次设计分割主要针对颜色和纹理来分割梅花图像，排除干扰。特征提取主要针对颜色特征，纹理特征，形状特征，以及局部区块特征。以及其他相关的图像处理操作，例如，填充，去噪，以及图像处理中算法otsu大津阈值分割，形态学等图像处理的操作。取得了较好的效果，为其他花卉的分类识别奠定了一定的基础。