基于无人机RGB图像颜色特征和形态特征的滴灌棉花苗期株数估算研究  

Estimation of the quantity of drip-irrigated cotton seedling based on color and morphological features of UAV captured RGB images

作　　者：蒋家乐 李容望 马学花 李明华 刘扬 芦阳 马富裕[1,2] Jiang Jiale;Li Rongwang;Ma Xuehua;Li Minghua;Liu Yang;Lu Yang;Ma Fuyu(School of Agriculture,Shihezi University/The Key Laboratory of Oasis Eco-Agriculture,Xinjiang Production and Construction Corps,Shihezi,Xinjiang 832003,China;National&Local Joint Engineering Research Center of Information Management and Application Technology for Modern Agricultural Production(Xinjiang Production and Construction Corps),Shihezi,Xinjiang 832003,China)

机构地区：[1]石河子大学农学院/新疆生产建设兵团绿洲生态农业重点实验室,新疆石河子832003 [2]现代农业生产信息化管理与应用技术国家地方联合工程研究中心(新疆兵团),新疆石河子832003

出　　处：《棉花学报》2022年第6期508-522,共15页Cotton Science

基　　金：国家自然科学基金(31860346);石河子大学青年创新人才培育计划项目(CXBJ202001)。

摘　　要：【目的】利用无人机图像数据的颜色特征和形态特征构建滴灌棉花苗期株数估算模型,为棉花田间精准管理提供理论依据。【方法】于2020―2021年开展试验,以鲁棉研24号为供试品种,设置3个不同种植密度,分别为:低密度(D1,6.9×10^(4)株·hm^(-2))、中密度(D2,13.8×10^(4)株·hm^(-2))、高密度(D3,24×10^(4)株·hm^(-2))。对出苗后25 d的无人机图像提取基于红绿蓝(red,green,and blue,RGB)的植被指数和目标形态特征,构建棉花株数估算特征集合;在自变量间的相关性分析的基础上,利用逐步多元回归的方法构建棉苗株数的估算模型,并进行验证。【结果】(1)三角绿度指数(triangular greenness index,TGI)、超绿指数(excess green index,ExG)、绿-蓝差值+修正超绿指数(green-blue difference+modified excess green index,GBDI+MExG)均对图像有较好的分割效果,其中TGI对棉花目标的分割完整度最高。(2)对比2种特征参数构建的棉花株数估算模型,基于目标形态特征的苗期棉花估算模型的拟合优度(R2=0.9355)要高于基于RGB植被指数的株数估算模型(R2=0.9036)。(3)基于RGB植被指数的株数估算模型在D1、D2、D3密度下估算精度分别为96.77%、99.55%和95.95%,整体估算精度为98.47%;基于目标形态特征的株数估算模型在D1、D2和D3密度下估算精度分别为99.98%、99.21%和97.92%,整体估算精度为99.21%。基于目标形态特征的株数估算模型的估算精度略高于基于RGB植被指数的株数估算模型,但2个模型在不同种植密度下均具有较好的估算效果。【结论】利用集成高分辨率传感器的低空无人机遥感平台,通过颜色特征和目标形态特征构建的滴灌棉花苗期株数估算模型均能有效、精准识别膜下滴灌棉花株数,可为后续棉花田间精准管理提供技术支撑。[Objective]A model for estimating the quantity of seedlings in drip-irrigated cotton using color characteristics and morphological characteristics of unmanned aerial vehicle(UAV)image data was constructed to provide a theoretical basis for accurate management in cotton field.[Methods]The experiment was carried out in 2020-2021 and the cultivar Lumianyan24 was used in the experiment.Three different planting densities were set as follow:low density(D1,6.9×10~4 plant·hm^(-2)),medium density(D2,13.8×10~4 plant·hm^(-2))and high density(D3,24×10~4 plant·hm^(-2)).The UAV images were obtained on the25 days old cotton seedlings,and the vegetation indices(VIs)of red,green,and blue(RGB)and target morphological features were extracted from the acquired UAV images.Based on the selected independent variable according to the correlation analysis,the model to estimate the quantity of cotton seedlings was constructed using stepwise multiple regression,followed by the model validation.[Results](1)Comparing the segmentation effects of extracting cotton targets by triangular greenness index(TGI),excess greenness index(ExG),and green-blue difference+modified excess greenness index(GBDI+MExG),all these three VIs had relatively good segmentation effects,while TGI showed the highest precision of segmentation of cotton targets.(2)Comparing the two cotton plant quantity estimation models constructed with the two feature parameters,the estimation model based on the target morphological features for cotton seedling(R~2=0.9355)is better than the estimation model based on the VI of RGB(R~2=0.9036).(3)The estimation accuracy of the VIs-based seedling quantity estimation model were 96.77%,99.55%,and 95.95%at D1,D2 and D3 densities respectively,and the overall estimation accuracy was 98.47%;the estimation accuracy of the plant estimation model based on the target morphological features at D1,D2 and D3 densities were 99.98%,99.21%,and97.92%respectively,and the overall estimation accuracy was 99.21%.The accuracy of the plant number estimatio

关 键 词：棉花 无人机 颜色特征 形态特征 株数估算 模型 

分 类 号：S562[农业科学—作物学] S127