| 基于小波变换压缩和支持向量机组的储粮害虫图像识别 |
| |
| 引用本文： | 廉飞宇,张元. 基于小波变换压缩和支持向量机组的储粮害虫图像识别[J]. 河南工业大学学报(自然科学版), 2006, 27(1): 21-25 |
| |
| 作者姓名： | 廉飞宇 张元 |
| |
| 作者单位： | 河南工业大学,信息科学与工程学院,河南,郑州,450052 |
| |
| 基金项目： | 河南工业大学引进人才专项基金(150162) |
| |
| 摘 要： | 使用小波变换对储粮害虫的高维图像矢量进行压缩,利用图像的高频部分对应于图像的边缘和轮廓,很好地压缩和表征了害虫图像的特征.提出了一种基于支持向量机(SVM)组的淘汰法.这种方法考虑到了各判别函数的VC置信范围的差异,同时利用判别函数间的冗余来降低识别误差.100幅害虫图像的识别结果表明,基于SVM的识别方法在识别效果、识别时间等方面都有显著的优越性.
|
| 关 键 词： | 储粮害虫 图像识别 小波变换 支持向量机 |
| 文章编号： | 1673-2383(2006)01-0021-04 |
| 修稿时间： | 2005-09-08 |
DETECTION OF PESTS IN STORED-GRAIN BASED ON IMAGE RECOGNITION |
| |
| LIAN Fei-yu,ZHANG Yuan. DETECTION OF PESTS IN STORED-GRAIN BASED ON IMAGE RECOGNITION[J]. Journal of Henan University of Technology Natural Science Edition, 2006, 27(1): 21-25 |
| |
| Authors: | LIAN Fei-yu ZHANG Yuan |
| |
| Abstract: | The high dimensional image vector of pests in stored grain was compressed with the wavelet transform.The high frequency part expresses the boundary and profile of the image,and the image is compressed and characterized effectively.Support Vector Machines(SVM) group incorporated with the elimination strategy is proposed to deal with the multi-class recognition problem.In this method, the difference among the confidence interval about the SVM is considered and the redundancy of each SVM was used to reduce the recognition error.The results of the recognition experiments with 100 images show that the method has reached a high recognition rate and a reasonable time. |
| |
| Keywords: | stored-grain pests image recognition wavelets transform support vector machines |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
