| 大豆病斑智能识别无损预处理及其特征提取方法的研究 |
| |
| 引用本文： | 关海鸥,黄燕.大豆病斑智能识别无损预处理及其特征提取方法的研究[J].河北农业大学学报,2010,33(5). |
| |
| 作者姓名： | 关海鸥 黄燕 |
| |
| 作者单位： | 1. 黑龙江八一农垦大学,信息技术学院,黑龙江大庆,163319
2. 黑龙江八一农垦大学,工程学院,黑龙江大庆,163319 |
| |
| 基金项目： | 智能化液体肥料施用系统的研究 |
| |
| 摘 要： | 以大豆病叶为例,针对在采集叶片图像过程中出现的几何失真问题,提出了基于投影模型的植物叶片校正算法,测量准确性只与校正方法有关,不受其他背景因素影响。在精确校正的基础上,又实现了叶片病斑区域的特征提取与计算,实验结果表明,该方法校正精度达99%,病斑区域提取精度达100%,为进一步的病害诊断奠定了先期基础。
|
| 关 键 词： | 大豆 投影模型 几何失真校正 病斑 特征提取 |
Study on the method of non-loss pre-processing and feature extraction for intelligent recognition of soybean diseased spots |
| |
| GUAN Hai-ou,HUANG Yan.Study on the method of non-loss pre-processing and feature extraction for intelligent recognition of soybean diseased spots[J].Journal of Agricultural University of Hebei,2010,33(5). |
| |
| Authors: | GUAN Hai-ou HUANG Yan |
| |
| Institution: | GUAN Hai-ou1,HUANG Yan2(1.College of Information Technology,Heilongjiang Bayi Agricultural University,Daqing 163319,China,2.College of Engineering,China) |
| |
| Abstract: | By using soybean as an example,this paper proposes a registration Algorithm named Projection Model,aiming to avoid Geometric distortion when collecting lamina images.The accuracy is related only to registration Algorithm.Furthermore,the diseased spots have been extracted and calculated,The results showed that the registration accuracy and extraction accuracy can reach 99% and 100% respectively.It is hoped this model provides a preliminary basis for disease diagnosis. |
| |
| Keywords: | Soybean projection model registration diseased spots feature extraction |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
