新疆农业科学 ›› 2019, Vol. 56 ›› Issue (7): 1312-1324.DOI: 10.6048/j.issn.1001-4330.2019.07.015

• 土壤肥料·贮藏加工·植物保护·畜牧水产 • 上一篇    下一篇

基于无人机影像天山云杉林主伐迹地提取研究

王雅佩, 王振锡, 刘梦婷, 李擎, 杨勇强   

  1. 新疆农业大学林学与园艺学院,新疆教育厅干旱区林业生态与产业技术重点实验室,乌鲁木齐 830052
  • 收稿日期:2019-04-12 出版日期:2019-07-20 发布日期:2019-09-05
  • 通信作者: 王振锡(1977- ),男,新疆喀什人,副教授,博士,硕士生导师,研究方向为林业3S技术及应用,(E-mail)wangzhenxi2003@163.com
  • 作者简介:王雅佩(1993-),女,新疆伊宁人,硕士研究生,研究方向为林业3S技术及应用,(E-mail)442413889@qq.com
  • 基金资助:
    新疆维吾尔自治区林业改革发展基金项目“新疆天保工程精准监测技术与评价体系研究”(XJTB20181102)

Research on Extraction of Final Felling Area of Picea Schr enkiana var tianshanica Based on UAV Image

WANG Ya-pei, WANG Zhen-xi, LIU Meng-ting, LI Qing, YANG Yong-qiang   

  1. Key Laboratory of Forestry Ecology and Industry Technology in Arid Region, Education Department of Xinjiang / College of Forestry and Horticulture,Xinjiang Agricultural University , Urumqi 830052, China
  • Received:2019-04-12 Online:2019-07-20 Published:2019-09-05
  • Correspondence author: WANG Zhen-xi(1977- ),male, native place: Xinjiang, professor, Research field: Forestry 3s technology and its application. (E-mail)wangzhenxi2003@163.com
  • Supported by:
    Supported by the Xinjiang Uygur Autonomous Region Forestry Reform and Development Fund Project "Study on Precision Monitoring Technology and Evaluation System of Natural Forest Conservation Project in Xinjiang" (XJTB20181102)

摘要: 【目的】通过分析天山云杉主伐迹地更新群落特征,将其划分为三个等级的主伐迹地,基于无人机影像,以实现提取不同等级主伐迹地的面积以及确定其分布,为新疆天保工程实施后山区森林资源更新恢复及评价提供科学依据。【方法】采用eCognition软件面向对象多尺度分割技术对研究区进行分割,通过最邻近分类以及SEaTH算法相结合的分类方法筛选分类特征,并建立类别层次结构及分类规则集,以实现提取不同等级主伐迹地面积,并进行精度评价。【结果】利用最邻近分类方法与SEaTH算法相结合的分类方法提取不同等级的主伐迹地,其总体分类精度达到81.82%,Kappa系数为0.74。提取主伐迹地Ⅰ、Ⅱ、Ⅲ的面积吻合度分别为87.09%、79.86%和66.33%;提取主伐迹地总面积为72.574 3 hm2,目视解译面积为92.174 9 hm2,面积相对误差为21.26%,面积吻合度为78.74%。【结论】该方法用于研究区提取主伐迹地信息是可行的。主伐迹地Ⅰ提取的面积吻合度最高,由于主伐迹地Ⅰ内的天山云杉林分株数密度较低,其林分平均冠幅、平均年龄等因子均相对较小,因此在影像上更好区分。主伐迹地Ⅱ分布面积最大,主伐迹地Ⅰ和Ⅲ分布面积较小,研究区内主伐迹地林分生长状况主要为主伐迹地Ⅱ状态

关键词: 天山云杉; 无人机影像; 主伐迹地; 面向对象; SEaTH算法

Abstract: 【Objective】 Based on the analysis of the characteristics of the regeneration community of the final felling area of Picea Schrenkiana var tianshanica in Tianshan Mountain, it was divided into three grades of final felling area. Based on UAV images, the area and distribution of the final felling area of different grades can be extracted and determined in the hope of providing technical support for the evaluation of forest resources renewal and restoration in mountainous areas after the implementation of Xinjiang natural forest conservation project to a certain extent. 【Method】The object-oriented multi-scale segmentation technology of eCognition software was used to segment the study area. Through the combination of nearest neighbor classification and SEaTH algorithm, the classification features were screened, and the classification hierarchy and classification rule set were established so as to realize the extraction of the area of the main cutting area of different grades, and carry out the precision evaluation.【Result】The nearest neighbor classification method combined with the SEaTH algorithm was used to extract the main logging sites of different grades. The overall classification accuracy was 81.82% and the Kappa coefficient was 0.74. The area coincidence of the final felling areaⅠ, Ⅱ and Ⅲ were 87.09%, 79.86% and 66.33%. The total area of the final felling area was 72.574,3 hm2, visual interpretation area was 92.174,9 hm2, area relative error was 21.26% and area coincidence was 78.74%.【Conclusion】It is feasible to use this method to extract the information of final felling area in the study area. Because of the lower number density of Picea Schrenkiana var tianshanica, the average crown width and average age of the spruce stand are relatively small, so it is better to distinguish in image. The distribution area of the main cutting site II is the largest, while the distribution area of the main cutting site Ⅰ and III is small. The stand growth status of the main cutting land in the study area is mainly in the second state of the cutting site.

Key words: Picea Schrenkiana var tianshanica; UAV image; final felling area; object-oriented; SEaTH algorithm 

中图分类号: 


ISSN 1001-4330 CN 65-1097/S
邮发代号:58-18
国外代号:BM3342
主管:新疆农业科学院
主办:新疆农业科学院 新疆农业大学 新疆农学会

出版单位:《新疆农业科学》编辑部
地址:乌鲁木齐市南昌路403号新疆农业科学院
邮编:830091
电话:0991-4502046
E-mail:xjnykx-h@xaas.ac.cn


版权所有 © 《新疆农业科学》编辑部
本系统由北京玛格泰克科技发展有限公司设计开发
总访问量: 今日访问: 在线人数:
网站
微信公众号
淘宝购买
微店购买