| 4种人工林下可燃物的计算机图像分类 |
| |
| 引用本文： | 金森,王亮.4种人工林下可燃物的计算机图像分类[J].东北林业大学学报,2003,31(5):1-3. |
| |
| 作者姓名： | 金森 王亮 |
| |
| 作者单位： | 1. 东北林业大学,哈尔滨,150040
2. 大连市林业勘察设计室 |
| |
| 基金项目： | 国家自然科学基金青年基金项目 (30 0 0 0 1 35)，哈尔滨市学科后备带头人基金项目资助 |
| |
| 摘 要： | 对水曲柳(Fraxinus mandshurica)、胡桃楸(Juglans mandshurica)、蒙古栎(Quercus mongolica)和落叶松(Larix gemlini)人工林下的可燃物的图像进行特征提取，建立了以距离为5个像素的自相关系数、距离为10个像素的自相关系数和行程长度为8的频数这三个特征为变量的线性分类器，将落叶松、蒙古栎、胡桃楸和水曲柳人工林下的可燃物较好地区分开来，分类准确率达91％，而且分类能力比较稳定。
|
| 关 键 词： | 人工林 可燃物 计算机图像 图像分类 水曲柳 胡桃楸 蒙古栎 落叶松 |
| 修稿时间： | 2003年4月5日 |
Classification of 4 Fuels under different Plantations by Their Computer Images |
| |
| Jin Sen,Wang Liang.Classification of 4 Fuels under different Plantations by Their Computer Images[J].Journal of Northeast Forestry University,2003,31(5):1-3. |
| |
| Authors: | Jin Sen Wang Liang |
| |
| Abstract: | Characteristics of images for four kinds of fuels under different plantations, namely, Manchurian ash, Manchurian walnut , Mongolian oak and larch, were extracted, including Euler number, autocorrelation coefficients at different distance and travel length frequency. A linear classification machine was established based on the autocorrelation coefficients at the distance of 5 and 10 pixels and the number of travel length of 8. The results indicate that the classification machine can group the four fuels well with 91% correctness and steady performance. |
| |
| Keywords: | Forest fuel Image Classification Pattern recognition |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
