论文部分内容阅读
遥感技术应用已成为中国中药资源普查的一个重要探索方向。以红花(Carthams Tinctorius L.)为种植型药用植物实验样本品种,分别基于分形理论和灰度共生矩阵(GLCM)两种方法提取不同的纹理特征,结合光谱信息对资源三号卫星(ZY-3)影像进行最大似然方法的监督分类,对比分析分类效果和精度评价。结果显示:加入纹理特征后,总体分类精度提高了0.49%—5.31%,Kappa系数提高了0.01—0.07,结合基于双毯法的分形纹理较GLCM纹理分类总体效果提高至少两倍,其中在Matlab环境下,使用5×5滑动窗口提取的分形纹理特征的分类效果最显著,总体分类精度提高了5.31%,Kappa系数提高了0.07。对于红花分类精度,引入分形纹理特征的分类精度提高到了100%,识别的红花样地效果最完整,破碎程度最小,与其他类别区分度最高;而引入GLCM的分类精度却降低了0.55%—1.28%,可见采用分形理论比采用GLCM提取纹理特征能够更加有效地辅助ZY-3影像识别种植型药用植物。
The application of remote sensing technology has become an important exploration direction of Chinese medicine resources census. Taking Carthams Tinctorius L. as test sample, the different texture features were extracted based on the fractal theory and GLCM, respectively. Based on the spectral information, ZY -3) images were supervised and classified by the method of maximum likelihood, and the classification and accuracy evaluation were comparatively analyzed. The results showed that the overall classification accuracy increased by 0.49% -5.31% and the Kappa coefficient increased by 0.01-0.07 after the texture features were added, and the fractal texture based on double-blanket method increased at least two times compared with the GLCM texture classification. In Matlab environment , The fractal texture feature extracted using 5 × 5 sliding window has the most obvious classification effect, the overall classification accuracy is increased by 5.31%, and the Kappa coefficient is increased by 0.07. For safflower classification accuracy, the classification accuracy introduced into fractal texture features increased to 100%, the safflower samples identified were the most complete, with the smallest degree of fragmentation, and the highest classification accuracy compared with other categories. However, the classification accuracy introduced by GLCM was reduced by 0.55% 1.28%, showing that using fractal theory than using GLCM to extract texture features can be more effective in assisting ZY-3 imaging identification of implanted medicinal plants.