基于手机相片的草地植被盖度估算方法研究

发布时间：2018-06-04 04:49

  本文选题：植被盖度 + 数字相片　； 参考：《浙江农业学报》2017年06期

【摘要】：选取6种基于RGB通道信息的植被指数(VEG、CIVE、EXG、EXGR、NGRDI、COM),借助自主开发的手机APP对6种方法展开对比研究,从草绿度、天气条件、盖度3个角度分析估算误差的变化规律,并从阈值随光照变化角度评估每种方法的稳定性。研究表明,6种方法估算精度均高于90%,其中,COM方法最高,达到95.41%,NGRDI方法估算精度最低,为92.87%。每种方法对深绿色草的估算误差均小于对黄绿色草,在阴天条件下(云量≥70%)的估算误差小于晴天条件下(云量≤10%)。盖度增加时,VEG、CIVE、EXG、COM方法估算误差增大,EXGR、NGRDI方法无明显变化规律。同日内不同时次,随着太阳高度角、光照强度的变化,6种方法阈值无明显变化(阈值波动≤0.02)。不同天气条件下,VEG、CIVE方法的阈值基本无变化(阈值波动≤0.01),其余方法变化明显(阈值波动≥0.03)。综上,6种方法均可满足在手机平台中应用的要求,COM方法精度最高,NGRDI方法精度最低。VEG、CIVE方法阈值设定无须考虑光照变化影响,较其他方法具有更好的通用性。
[Abstract]:Six kinds of vegetation index based on RGB channel information were selected to analyze the variation of estimation error from three angles of grass green degree, weather condition and coverage degree, with the help of self-developed mobile phone APP. The stability of each method was evaluated from the point of view of threshold changing with light. The results show that the estimation accuracy of the six methods is higher than 90, among which the com method is the highest, and the precision of NGRDI method is the lowest, 92.87. The estimation error of each method for dark green grass is lower than that for yellowish green grass, and the estimation error for dark green grass is less than that for sunny day (cloud amount 鈮，

本文编号：1976009