摘要
[目的]通过计算小麦的绝对高度来监测小麦倒伏状况,基于无人机影像获取小麦倒伏面积及其分布,为快速进行农业灾害评估提供科学依据。[方法]对奇台县境内不同地形部位的小麦地块进行调查和研究,结合小麦高度的实测值构建倒伏模型,以无人机平台获取的DSM影像为数据源,通过DSM差减法得到小麦实际高度,选取3个典型倒伏地块,用构建的倒伏模型对研究区进行分类,用像元统计的方法提取倒伏面积并采用混淆矩阵的方法对模型的分类精度进行验证。[结果]倒伏主要发生在沙漠边缘、中部平原以及丘陵区的灌溉耕地,南部山区的旱耕地未发生倒伏;模型解译的总体分类精度均在83.33%以上,Kappa系数在0.75以上,表明用文中建立的倒伏模型对奇台县灌浆期小麦进行倒伏监测是可行的。[结论]利用无人机获取的DSM数据可以有效监测小麦倒伏信息。
[Objective]The absolute height of wheat was calculated to monitor the situation of wheat lodging, and the area and distribution of wheat lodging were obtained based on UAV images, providing scientific basis for rapid agricultural disaster assessment.[Method]The investigation of the wheat plots in different terrains in Qitai County was conducted. The lodging model was constructed based on the measured height of wheat. DSM image acquired by UAV platform is used as base data, DSM subtraction to get the actual height of the wheat. Three typical fields were selected and the lodging model was used to classify the study area. The method of pixel statistics is used to extract the lodging area and the method of confounding matrix is used to verify the classification accuracy of the model.[Result]The results showed that lodging mainly occurred in the irrigated land of the desert edge, the central plain and the hilly area, while the dry land in the southern mountainous area without lodging. The overall classification accuracy of the model interpretation is above 83.33%, and the Kappa coefficient is above 0.75. It shows that the lodging model is feasible to monitor the lodging of wheat during the filling period in Qitai County.[Conclusion]The above results show that DSM data obtained by UAV can effectively monitor wheat lodging information.
作者
胡琪
武红旗
轩俊伟
范燕敏
王德俊
谷海斌
HU Qi;WU Hong-qi;XUAN Jun-wei(College of Grassland and Environment Science,Xinjiang Agricultural University,Urumqi,Xinjiang 830052;Xinjiang Key Laboratory of Soil and Plant Ecological Process,Urumqi,Xinjiang 830052)
出处
《安徽农业科学》
CAS
2020年第16期227-231,共5页
Journal of Anhui Agricultural Sciences
基金
国家自然科学基金项目(31560340)。
