当前位置: 首页 > 文章 > 基于无人机倾斜摄影测量的树冠体积及表面积提取算法对比分析 林业工程学报 2022 (3) 166-173
Position: Home > Articles > Comparative analysis of extraction algorithms for crown volume and surface area using UAV tilt photogrammetry Journal of Forestry Engineering 2022 (3) 166-173

基于无人机倾斜摄影测量的树冠体积及表面积提取算法对比分析

作  者:
王玉堂;王佳;牛利伟;常书萍;孙露
单  位:
北京林业大学林学院;北京林业大学精准林业北京市重点实验室
关键词:
无人机倾斜摄影测量;树冠体积;树冠表面积;点云密度;冠型
摘  要:
树冠体积和表面积是树冠结构的重要特征因子,是准确监测树木生长状况的关键参数。利用无人机倾斜摄影测量可以快速获取树冠三维结构,因此对比分析树冠体积和表面积的不同提取算法,可以提高林业调查中树冠体积和表面积的提取效率和精度。通过无人机倾斜摄影测量获取树木三维点云数据,分别使用体元法和数字高程法计算树冠体积和表面积,分析两种算法在不同树种间计算结果的差异。研究结果表明,利用数字高程法得到的树冠体积和表面积明显大于体元法,分别将点云密度和冠型参数与两种算法计算结果的相对差值进行拟合,拟合结果表明:点云密度只对树冠体积的计算有影响,决定系数R~2为0.443 1,冠型参数对树冠体积和表面积的计算均有影响,决定系数R~2分别为0.349 6和0.100 1。对于精度要求较高的研究,体元法更加准确,而对于树冠内部空隙少且树冠表面较规则的树木,在计算树冠体积和表面积时,可以用数字高程法代替体元法来提高工作效率。
译  名:
Comparative analysis of extraction algorithms for crown volume and surface area using UAV tilt photogrammetry
作  者:
WANG Yutang;WANG Jia;NIU Liwei;CHANG Shuping;SUN Lu;College of Forestry, Beijing Forestry University;Beijing Key Laboratory of Precision Forestry,Beijing Forestry University;
关键词:
UAV tilt photogrammetry;;crown volume;;crown surface area;;point cloud density;;crown type
摘  要:
Canopy volume and surface area are important characteristic factors of canopy structure, which are key parameters for accurately monitoring tree growth. The use of drone tilt photogrammetry can quickly examine the three-dimensional structure of the crown. Therefore, the comparison and analysis of different extraction algorithms for crown volume and surface area can improve the extraction efficiency and accuracy of the crown volume and surface area in forestry investigations. In this study, the main street trees in Huairou District of Beijing were used as the research objects. The 3 D point cloud data of trees were obtained through the UAV tilt photogrammetry, the crown volume and surface area were calculated using the voxel method and digital elevation method, respectively, and the difference in settlement results between the two algorithms between different trees species was analyzed. The research results showed that the voxel method reflected the internal structure of the canopy, while the digital elevation method discovered that there was no void inside the canopy. Therefore, the volume and surface area of the crown obtained by the digital elevation method were significantly larger than that of the voxel method. The point cloud density and crown parameters were respectively fitted to the relative difference between the calculation results of the two algorithms. The fitting results showed that the point cloud density only affected the crown volume with a correlation coefficient R~2 of 0.443 1. While the crown parameters influenced the calculation of crown volume and surface area, with the correlation coefficients R~2 of 0.349 6 and 0.100 1, respectively. Therefore, in practical applications, although the voxel method can reflect the accuracy more realistically from the theory of the two algorithms, the calculation process is slow. In addition, it is relatively troublesome to develop the calculation tools by yourself. However, the digital elevation method can be used. Rely on the volume surface area calculation tool provided by the existing software, the crown structure factor can be quickly obtained. Therefore, for trees with small internal space and relatively regular crowns, such as arborvitae, or for studies that do not consider the internal structure of trees, such as three-dimensional green quantity, the digital elevation method still has a good application prospect. If the calculation results are required to be accurate, or there are many gaps in the canopy, the voxel method is more suitable.

相似文章

计量
文章访问数: 13
HTML全文浏览量: 0
PDF下载量: 0

所属期刊

推荐期刊