| 图像处理技术在赤潮生物自动识别中的应用 |
| |
| 引用本文： | 王学民,张磊,王明时.图像处理技术在赤潮生物自动识别中的应用[J].计算机辅助设计与图形学学报,2004,16(2):158-162. |
| |
| 作者姓名： | 王学民 张磊 王明时 |
| |
| 作者单位： | 天津大学精密仪器与光电子工程学院,天津,300072 |
| |
| 基金项目： | 国家“八六三”高技术研究发展计划 (2 0 0 2AA92 63 60 )资助 |
| |
| 摘 要： | 根据赤潮生物显微图像的特点，研究图像的预处理、分割和特征提取的处理过程．通过与其他自动分割方法进行比较，证明采用二维最大熵法可以有效地进行赤潮生物图像的自动阈值分割，并结合轮廓分析法和纹理分析法对于赤潮生物图像进行特征提取．实验结论表明，与人工识别计数方式比对，识别符合率大于80％。
|
| 关 键 词： | 图像处理 赤潮生物 生物自动识别 特征提取 图像分割 二维最大熵法 数学形态学 赤潮监测 |
Image Processing for Automated Identification of Harmful Algae |
| |
| Wang Xuemin,Zhang Lei,Wang Mingshi.Image Processing for Automated Identification of Harmful Algae[J].Journal of Computer-Aided Design & Computer Graphics,2004,16(2):158-162. |
| |
| Authors: | Wang Xuemin Zhang Lei Wang Mingshi |
| |
| Abstract: | The pre-processing, segmentation and feature extraction of harmful algae from micro-image are presented. Theoretical analysis and experiment results show that 2D entropic threshold algorithm is capable of segmenting the harmful algae's image effectively, as compared with other auto-segmentation methods. Mathematical morphology was applied to increase the veracity of feature extraction. Feature extraction and classification are accomplished by using the image edge and texture. Experiment result indicates that the classification accuracy is over 80%, as compared with the artificial identification. |
| |
| Keywords: | harmful algae image processing mathematical morphology image segmentation |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
