| 基于支持向量机的栗属树种分类研究 |
| |
| 引用本文： | 林大辉,陈秋妹,宁正元.基于支持向量机的栗属树种分类研究[J].莆田学院学报,2009,16(5):39-42,46. |
| |
| 作者姓名： | 林大辉 陈秋妹 宁正元 |
| |
| 作者单位： | 福建农林大学计算机与信息学院,福建,福州,350002 |
| |
| 基金项目： | 福建省自然科学基金资助项目，福建省教育厅资助项目 |
| |
| 摘 要： | 形状特征是物体识别的重要依据。同时,分类算法的选择也将对识别的性能造成很大影响。围绕上述两个问题,以栗属树种的果实图像为例,在准确分割出目标图像的基础上,分别应用不变矩和边界矩提取其形状特征值,并使用支持向量机算法对栗属树种果实图像进行分类。实验结果表明:基于支持向量机的栗属树种果实图像分类识别准确率可达到87.5%,识别的结果较为理想。
|
| 关 键 词： | 支持向量机 不变矩 边界矩 树种分类 |
Study on Support Vector Machine Based Chestnut Species Classification |
| |
| LIN Da-hui,CHEN Qiu-mei,NING Zheng-yuan.Study on Support Vector Machine Based Chestnut Species Classification[J].journal of putian university,2009,16(5):39-42,46. |
| |
| Authors: | LIN Da-hui CHEN Qiu-mei NING Zheng-yuan |
| |
| Abstract: | Shape is important for object recognition.At the same time,the alternative of classification algorithms will have a great impact on accuracy of the object identification.As a result,taking chestnut fruit image as an example,this papers,on the basis of accurate segmentation of the target image,applies invariant moments and boundary moments to extract features of the shape respectively,then according to the extracted shape features,it applies support vector machine to classify the images of the chestnut fruit.The experimental results show that: based on support vector machine,classification accuracy of chestnut fruit image can reach 87.5%,and the performance of classification is good. |
| |
| Keywords: | support vector machine invariant moments boundary moments species classification |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
