本文選題：無人機(jī)　切入點(diǎn)：遙感　出處：《農(nóng)業(yè)工程學(xué)報(bào)》2017年19期 　論文類型：期刊論文

【摘要】：植株數(shù)量識別是油菜機(jī)械直播效果和質(zhì)量評估的關(guān)鍵。該文針對油菜機(jī)械直播田間植株數(shù)量檢測中人工統(tǒng)計(jì)耗時(shí)、費(fèi)工、效率低下的現(xiàn)實(shí),通過自主搭建的低空無人機(jī)遙感平臺采集油菜機(jī)械直播區(qū)域的遙感影像,基于超高分辨率(0.18 cm/pixel)遙感影像計(jì)算的顏色植被指數(shù)進(jìn)行油菜目標(biāo)識別及其形態(tài)特征信息提取。結(jié)合田間調(diào)查數(shù)據(jù),采用逐步回歸分析方法,建立了機(jī)械直播油菜在苗期的株數(shù)與遙感特征信息之間的關(guān)系。結(jié)果顯示,油菜目標(biāo)的株數(shù)與其外接矩形的長寬比、像素分布密度和周長柵格數(shù)具有較好的線性關(guān)系,回歸模型的決定系數(shù)R~2為0.803,并通過顯著性檢驗(yàn),其標(biāo)準(zhǔn)估計(jì)誤差為0.699。模型檢驗(yàn)結(jié)果顯示,觀測值與預(yù)測值之間的R~2為0.809,均方根誤差RMSE為0.728。研究結(jié)果表明,利用集成超高分辨率傳感器的低空無人機(jī)遙感平臺,通過計(jì)算顏色植被指數(shù)并分析油菜目標(biāo)數(shù)量與形態(tài)特征的相關(guān)性,能有效識別油菜機(jī)械直播的出苗株數(shù),可為后續(xù)油菜機(jī)械直播效果的快速、準(zhǔn)確評估提供技術(shù)支持。
[Abstract]:The identification of plant quantity is the key to the evaluation of the effect and quality of direct seeding of rapeseed mechanically. This paper aims at the fact that the manual statistics of the number of plants in the field of direct seeding of rapeseed is time-consuming, laborious and inefficient. Remote sensing images of rapeseed machinery direct seeding area were collected by self-built low altitude UAV remote sensing platform. Based on the color-vegetation index calculated from the super-high resolution 0.18 cm / pixel remote sensing image, rapeseed target recognition and morphological information extraction were carried out, and the stepwise regression analysis method was used to combine the field investigation data. The relationship between the number of plants in seedling stage and remote sensing information was established. The results showed that there was a good linear relationship between the number of plants and the ratio of length to width of the outer rectangle, the distribution density of pixels and the number of girth grids. The determination coefficient of regression model Ru 2 is 0.803, and the standard estimation error is 0.699 through significance test. The results of model test show that the R2 between the observed value and the predicted value is 0.809, and the root mean square error (RMSE) is 0.728. Based on the remote sensing platform of low altitude UAV integrated with ultra-high resolution sensor, by calculating the color vegetation index and analyzing the correlation between the number of rapeseed targets and the morphological characteristics, the number of emergence plants of rapeseed mechanically direct seeding can be effectively identified. It can provide technical support for the rapid and accurate evaluation of the direct seeding effect of rapeseed.
【作者單位】： 華中農(nóng)業(yè)大學(xué)資源與環(huán)境學(xué)院;農(nóng)業(yè)部長江中下游耕地保育重點(diǎn)實(shí)驗(yàn)室;華中農(nóng)業(yè)大學(xué)工學(xué)院;中國科學(xué)院測量與地球物理研究所;華中農(nóng)業(yè)大學(xué)理學(xué)院;
【基金】：國家自然科學(xué)基金(41201364,31501222) 中央高�；究蒲袠I(yè)務(wù)費(fèi)專項(xiàng)資金資助(2017JC038,2015BQ026,2014QC013)
【分類號】：S233.75;TP751
，

本文編號：1600255