基于无人机高光谱影像的火龙果波段筛选研究  

Study on band screening of pitaya fruit based on unmanned aerial vehicle hyperspectral image

作　　者：陈智虎 舒田[1] 刘春艳 许元红 CHEN Zhi-hu;SHU Tian;LIU Chun-yan;XU Yuan-hong(Guizhou Institute of Agricultural Science and Technology Information,Huaxi 550006,Guizhou,China)

机构地区：[1]贵州省农业科技信息研究所,贵州花溪550006

出　　处：《湖北农业科学》2022年第22期158-162,共5页Hubei Agricultural Sciences

基　　金：贵州省科学技术基金项目[黔科合基础-ZK(2021)一般130号];贵州省科技计划项目[黔科合服企(2021)];贵州省农业科学院青年基金项目[黔农科院青年基金(2019)19号]。

摘　　要：为解决卫星高光谱影像分辨率低、获取周期长、价格昂贵等问题,利用无人机搭载高光谱传感器能够获取分辨率高、价格较低且时效性较好的优势,以关岭县上关镇乐安村火龙果(Hylocereus undatus‘Foo-Lon’)种植示范园区为研究区域,使用六旋翼无人机M600搭载PikaXC2传感器获取高分辨率的高光谱影像数据。结果表明,通过对高光谱影像波段标准差和相关性进行分析,初步筛选出适合火龙果分析研究的波段185个,其中可见光波段109个、近红外波段76个,波段相关性系数的分布呈现中间高、两边低的特点。可见光波段相关性系数均值均大于近红外波段内部之间的均值。In order to solve the problems of low resolution,long acquisition period and high price of satellite hyperspectral image,the advantages of high resolution,low price and good timeliness could be obtained by using the hyperspectral sensor mounted on the UAV.Taking the Dragon Fruit Planting Demonstration Park in Lean Village,Shangguan Town,Guanling County as the research area,the six-rotor UAV M600 mounted with the PikaxC2 sensor was used to obtain high-resolution hyperspectral image data.By analyzing the band standard deviation and correlation of hyperspectral images,185 bands suitable for the analysis and research of pitaya fruit were selected preliminarily.Among them,109 were visible bands and 76 were near-infrared bands.The distribution of correlation coefficients of bands was high in the middle and low on both sides.The mean values of correlation coefficients in visible light bands were greater than those within the near-infrared bands.

关 键 词：高光谱影像 波段 无人机 火龙果(Hylocereus undatus‘Foo-Lon’) 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置] S667.9[自动化与计算机技术—控制科学与工程]