高级检索
    孙钊, 潘磊, 孙玉军. 基于无人机影像的高郁闭度杉木纯林树冠参数提取[J]. 北京林业大学学报, 2020, 42(10): 20-26. DOI: 10.12171/j.1000-1522.20190386
    引用本文: 孙钊, 潘磊, 孙玉军. 基于无人机影像的高郁闭度杉木纯林树冠参数提取[J]. 北京林业大学学报, 2020, 42(10): 20-26. DOI: 10.12171/j.1000-1522.20190386
    Sun Zhao, Pan Lei, Sun Yujun. Extraction of tree crown parameters from high-density pure Chinese fir plantations based on UAV images[J]. Journal of Beijing Forestry University, 2020, 42(10): 20-26. DOI: 10.12171/j.1000-1522.20190386
    Citation: Sun Zhao, Pan Lei, Sun Yujun. Extraction of tree crown parameters from high-density pure Chinese fir plantations based on UAV images[J]. Journal of Beijing Forestry University, 2020, 42(10): 20-26. DOI: 10.12171/j.1000-1522.20190386

    基于无人机影像的高郁闭度杉木纯林树冠参数提取

    Extraction of tree crown parameters from high-density pure Chinese fir plantations based on UAV images

    • 摘要:
        目的  冠幅是树冠结构的重要特征因子,直接影响树木的生产力和生命力,郁闭度是反映森林冠层结构与密度以及评价森林经营管理采伐强度的重要指标之一。利用无人机可以云下飞行,易于获取图像,精度高,低成本等优势,研究无人机影像上提取树冠参数的方法,使无人机影像提取林木树冠参数的操作系统化,实现精准高效的森林资源清查和监测。
        方法  以福建将乐林场杉木人工纯林为研究对象,采用四旋翼无人机影像为数据源,基于面向对象分类的方法,将杉木纯林的树冠参数从无人机影像中提取出来。面向对象分类的方法需要先利用ESP工具选取最优分割尺度,然后根据影像的分割结果将树冠对象聚为一类,进而统计每个树冠对象栅格像素个数计算出树冠冠幅面积以及林分郁闭度。
        结果  面向对象分类有效地对高郁闭度林分进行了树冠的提取。在分割尺度为70时,单木树冠分割效果最好,树冠被单独分割出来,但也存在一定的过分割以及未分割的问题,以至于部分单木的丢失。分割结束后,对分割对象进行特征空间的优化,选取适当的分类特征,最终将研究区分为树冠和林隙两类。通过统计每个对象栅格点数,计算得出的林分因子包括林分郁闭度,树冠面积。以地面实测数据作为参考,冠幅面积提取精度为0.829 1,林分郁闭度测量精度为0.973 1。
        结论  研究结果表明,基于无人机高分辨率影像的树冠参数提取在高郁闭度林分同样适用,能有效提高森林资源调查的效率并且能够满足森林资源调查的精度。

       

      Abstract:
        Objective  Crown width is an important characteristic factor of canopy structure, which directly affects the productivity and vitality of trees. The forest canopy density is one of the important indexes to reflect forest canopy structure and density and to evaluate forest management and logging intensity. UAV has the advantages of easily getting high-resolution remote sensing images with high precision and low cost. Studying the method of extracting canopy parameters using UAV images is of great significance for improving the accuracy and efficiency of forest resource inventory and monitoring.
        Method  Taking Chinese fir plantation in Jiangle Forest Farm of Fujian Province, eastern China as the research object, using the quadrotor UAV CCD image data as the data source, based on the object-oriented classification method, the canopy parameters of the Chinese fir plantation were extracted from the UAV images. Then the canopy objects were grouped into one group according to the segmentation results of the images, and the number of raster pixels of each canopy object was counted to calculate the canopy width area and canopy density.
        Result  The object-oriented classification effectively extracted the crown of high canopy density stand. When the segmentation scale was 70, the segmentation of single tree had the best effect. Some single trees were lost during the segmentation process because of over-segmentation and under-segmentation. After completing the segmentation, optimizing the feature space of the segmented object and selecting appropriate classification features, finally the study area was divided into two types: canopy and forest gap. By counting the number of grid points of each object, the calculated stand factors included canopy density and crown area. With the measured data on the ground as reference, the crown area extraction accuracy was 0.829 1, and the forest canopy density measurement accuracy was 0.973 1.
        Conclusion  The results show that the canopy parameter extraction based on high-resolution image of UAV is also applicable in high-canopy closed forest stands, which can effectively improve the efficiency and accuracy of forest resource survey.

       

    /

    返回文章
    返回