| 基于可见光光谱的幼龄沉香图像分割与特征提取研究 |
| |
| 引用本文： | 王鹏,王雪峰.基于可见光光谱的幼龄沉香图像分割与特征提取研究[J].西北林学院学报,2023,38(2):217-222. |
| |
| 作者姓名： | 王鹏 王雪峰 |
| |
| 作者单位： | （中国林业科学研究院 资源信息研究所，北京 100091） |
| |
| 摘 要： | 为促进数字图像处理技术在珍贵树种营养分析中的高效应用，以幼龄沉香为研究对象，运用大津法与K-Means算法分别对试验获取的幼龄沉香可见光图像进行分割，对2种分割算法进行比较研究。基于图像分割结果，提取R、G、B等8种颜色特征并进行主成分分析，同时计算沉香图像的最小外接矩形的矩形度RE。结果表明，1）大津法与K-Means算法均可实现对多张幼龄沉香可见光图像的分割，大津法较K-Means算法分割速度快，但分割精度小于K-Means算法，在具体分割时应根据实际需要对2种算法进行选择。2）提取的8种颜色特征的3个主成分累计贡献率可达到99%，可作为颜色特征；最小外接矩形的矩形度RE能够表达沉香轮廓内面积CA与最小外接矩形面积LA的比值，可作为形状特征，将这种特征因子用于构建沉香微量元素含量预测模型，有利于缩短建模时间并提高模型的精度。综上所述，研究结果可促进数字图像处理技术在珍贵树种营养诊断中的进一步发展，为精准林业提供参考。
|
| 关 键 词： | 幼龄檀香 可见光图像 图像分割 特征提取 |
Image Segmentation and Feature Extraction of Juvenile Agarwood Based on Visible Light Spectrum |
| |
| WANG Peng,WANG Xue-feng. Image Segmentation and Feature Extraction of Juvenile Agarwood Based on Visible Light Spectrum[J].Journal of Northwest Forestry University,2023,38(2):217-222. |
| |
| Authors: | WANG Peng WANG Xue-feng |
| |
| Affiliation: | (Institute of Forest Resource Information Techniques,Beijing 100091,China) |
| |
| Abstract: | In order to promote the efficient application of digital image processing technology in the nutritional analysis of precious tree species,this study took young agarwood as the research object,and used the best Otsu method in threshold segmentation and the best K-Means algorithm in cluster segmentation.The visible light image of juvenile agarwood obtained in the experiment was segmented,and the two segmentation algorithms were compared based on the manually segmented image.Based on the results of image segmentation,eight color features such as R,G,and B were extracted and principal component analysis was performed.At the same time,the rectangle degree RE of the smallest bounding rectangle of the agarwood image was calculated.The research results showed that 1) both the Otsu method and the K-Means algorithm could realize the segmentation of multiple visible light images of young agarwood.The Otsu method had faster segmentation speed than the K-Means algorithm,but the segmentation accuracy was lower than that of the K-Means algorithm.The two algorithms should be selected according to actual needs.2) The cumulative contribution rate of the three principal components of the extracted 8 color features could reach 99%,which could be used as color features; the rectangle degree RE of the minimum circumscribed rectangle could express the ratio of the inner area CA of the agarwood outline to the minimum circumscribed rectangle area LA.It could be used as a shape feature,and this feature factor could be used to build a prediction model for the content of trace elements in agarwood,which was beneficial to shorten the modeling time and improve the accuracy of the model.In conclusion,this study can promote the further development of digital image processing technology in the nutritional diagnosis of precious tree species,and provide a reference for precision forestry. |
| |
| Keywords: | young sandalwood visible light image image segmentation feature extraction |
|
| 点击此处可从《西北林学院学报》浏览原始摘要信息 |
|
点击此处可从《西北林学院学报》下载全文 |
|
