• 首页关于本刊期刊订阅编委会作者指南过刊浏览
高大中,林海,林乐乐,崔国发.2021.利用小型无人机监测西洞庭湖 水鸟的可行性探讨.动物学杂志,56(1):100-110.
利用小型无人机监测西洞庭湖 水鸟的可行性探讨
The Feasibility of Wetland Waterfowl Monitoring Method Based on UAV Remote Sensing
投稿时间:2020-03-29  修订日期:2021-01-04
DOI:10.13859/j.cjz.202101012
中文关键词:  无人机  湿地  水鸟  监测  越冬地
英文关键词:Unmanned aerial vehicle  Wetland  Waterfowl  Monitoring  Wintering grounds
基金项目:国家级大学生创新创业训练计划项目(No. G201910022068)
作者单位E-mail
高大中 北京林业大学生态与自然保护学院 北京 100083 gaodazhong@bjfu.edu.cn 
林海 北京林业大学生态与自然保护学院 北京 100083 850682892@qq.com 
林乐乐 北京林业大学生态与自然保护学院 北京 100083 linlele@bjfu.edu.cn 
崔国发 北京林业大学生态与自然保护学院 北京 100083 fa6716@163.com 
摘要点击次数: 333
全文下载次数: 352
中文摘要:
      传统鸟类监测方法具有调查时间长、人力物力消耗大、调查结果不准确的局限。近年来,无人机遥感技术在生态学领域的应用日益广泛,但在鸟类调查上仍缺乏成熟的技术方法。本研究于2019年11月19至25日期间,使用搭载可见光相机的小型多旋翼无人机大疆Mavic 2行业变焦版,在湖南西洞庭湖国家级自然保护区内的四个水鸟集群分布区域划定监测样区,规划航线,设定合适的飞行和拍摄参数后采集遥感数据。根据水鸟对无人机的反应程度划分不同的惊扰等级,记录拍摄过程中鸟类受惊扰情况。利用图像拼接软件PTGui Pro 11.0,对采集到的影像进行拼接、匀色等解译预处理操作。对合成后的遥感影像建立水鸟分类标注表,进行人工解译,并对调查过程中的水鸟惊扰情况进行统计分析。本研究共进行11次飞行调查,获取10个样区数据,最大样区面积约为18 hm2,75 m飞行高度下影像分辨率为0.012 m/像素,对样区内6种体型较大的水鸟——苍鹭(Ardea cinerea)、大白鹭(A. alba)、小天鹅(Cygnus columbianus)、凤头麦鸡(Vanellus vanellus)、绿翅鸭(Anas crecca)和罗纹鸭(Mareca falcata)进行了分类和计数。绿翅鸭和罗纹鸭二者依靠遥感图像无法区分,其余4种拍摄到的水鸟均成功解译和计数。惊扰等级记录显示,本次无人机调查对水鸟的惊扰程度较弱。结果表明,基于搭载可见光相机的小型无人机对湿地大型和中型水鸟进行快速遥感调查监测具有一定的可行性,在湖泊湿地类型的鸟类调查中具有应用潜力;通过选择合适的飞行平台,设定适当的飞行高度、飞行速度和图像重叠度等参数,能够在保证解译结果准确性的同时,避免对水鸟的过度干扰。
英文摘要:
      Traditional bird monitoring methods have the limitations of long investigation time, large consumption of manpower and material resources, as well as inaccurate identifications. In recent years, unmanned aerial vehicle (UAV) remote sensing technology has been widely used in the field of ecology, however, it is still a lack of developed methods for its application in bird survey. In this study, a micro-UAV (DJI Mavic 2) equipped with a visible light camera was used to conduct a waterfowl survey in the West Dongting Lake National Nature Reserve in Hunan Province. Four survey areas were selected inside the reserve where the target species were concentrated. With flight routes in each area planned and the flight and shooting parameters set appropriately, remote sensing data were collected (Fig. 1). Meanwhile, the birds’ response to the disturbance caused by the shooting process was ranked and recorded (Table 1). Subsequently, the image splicing software PTGui Pro 11.0 was employed to splice the collected images and adjusted their color balance. Finally, based on the synthesized remote sensing image, the classification and annotation table of waterfowl was established and the waterfowl in the image were manually interpreted. Also, the disturbance ranks of the investigation to the waterfowl was analyzed statistically. We conducted 11 flights in this study, of which the largest flight sites had an area of 18 hm2 and the resolution of 0.012 m/pixel at the flight altitude of 75 m. We successfully obtained data from ten survey flights. We tried to identify and count, in our obtained images, six main waterfowl species with relatively large body size: Ardea alba, A. cinerea, Anas crecca, Mareca falcata, Cygnus columbianus and Vanellus vanellus (Table 2, Fig. 2). The results showed that the four target species can be identified and counted accurately with the remote sensing images, however, the A. crecca and M. falcata could not be distinguished from each other. Alarm rating records show that the drone investigation had a minor impact on waterfowl. In conclusion, it is feasible to conduct rapid remote sensing monitoring of large and medium sized waterfowl in wetland based on micro-UAV equipped with visible light cameras. In order to achieve relatively higher accuracy of interpretation and prevent excessive interference to waterfowl at the same time, parameters such as flight altitude, flight speed and image overlap need to be set appropriately.
附件
查看全文  查看/发表评论  下载PDF阅读器