%0 Journal Article
%T 基于低空遥感的若尔盖高寒草地鼠害监测研究<br>Research on Monitoring Rodent Pests in Ruoergai Alpine Grassland Based on Low-Altitude Remote Sensing
%A 樊松林
%A 仙巍
%A 杨增平
%J International Journal of Ecology
%P 1-13
%@ 2324-7975
%D 2024
%I Hans Publishing
%R 10.12677/IJE.2024.131001
%X 若尔盖草原作为我国最大的高寒草原之一，具有重要的生态和经济价值。而草原鼠害是我国草原的主要生物灾害之一，其威胁着若尔盖草原畜牧业的可持续发展，为了及时有效地监测和应对草原鼠害问题，本研究基于多源遥感数据，探讨了若尔盖草原鼠害监测的方法。首先利用无人机影像发现并提取出鼠害信息，同时，结合气象数据和地面调查数据，选取了归一化植被指数(Normalized Difference Vegetation Index, NDVI)、坡向、气温、降水、土地利用等影响因子对若尔盖县的鼠害情况进行分析与探讨，采用随机森林模型模拟若尔盖县鼠害分布状况，并分析其空间分布特征。研究结果表明：利用随机森林对鼠害进行模拟，准确率可达到81.8%，kappa系数为0.748，结果与地面调查数据较为一致，具体一下，若尔盖草原东南部地区鼠害影响范围较小，往北则鼠害趋势明显加重，呈现多尺度密集分布的特点。本文可为高原地区鼠害监测研究提供参考，并具有较好的可操作性。<br />
Ruoergai Grassland, as one of China’s largest alpine grasslands, holds significant ecological and economic value. However, rodent infestation is a major biological disaster in Chinese grasslands, posing a threat to the sustainable development of animal husbandry in Ruoergai grassland. In or-der to timely and effectively monitor and address rodent infestation issues, this study explores methods for monitoring rodent infestation in Ruoergai grassland using multisource remote sensing data. First, rodent information is discovered and extracted from drone images. In addition, meteorological data and ground survey data are combined, and factors such as Normalized Dif-ference Vegetation Index (NDVI), aspect, temperature, precipitation, and land use are analyzed and discussed in relation to rodent infestation in Ruoergai County. A random forest model is employed to simulate the distribution of rodent infestation in Ruoergai County and analyze its spatial distribution characteristics. The research results indicate that using the random forest model for rodent infestation simulation achieves an accuracy of 81.8%, with a kappa coefficient of 0.748, which is consistent with ground survey data. Specifically, the southeastern part of Ruoergai Grassland shows a smaller impact from rodent infestation, while moving north, the rodent infes-tation trend significantly worsens, displaying a multi-scale dense distribution pattern. This study can provide a reference for rodent infestation monitoring research in high-altitude regions and is operationally feasible.
%K 若尔盖草原，高原鼠害，无人机，多源遥感数据，随机森林<br>Ruoergai Grassland
%K Highland Rodent Pests
%K Unmanned Aerial Vehicle
%K Multi-Source Remote Sensing Data
%K Random Forest
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=80511