本文關(guān)鍵詞： 植被覆蓋 土壤 水分含量 土層深度 高光譜遙感　出處：《土壤通報(bào)》2016年02期 　論文類型：期刊論文

【摘要】：在棉花大田水分試驗(yàn)的基礎(chǔ)上,采用自主設(shè)計(jì)的不同土層取樣方法,同步獲取了棉花冠層高光譜數(shù)據(jù)和不同深度土壤的水分含量數(shù)據(jù)以及棉花冠層水分含量數(shù)據(jù),分析了棉花冠層含水量與土壤含水量之間的關(guān)系、棉花冠層高光譜數(shù)據(jù)與土壤含水量之間的相關(guān)性,構(gòu)建了基于棉花冠層高光譜數(shù)據(jù)的土壤水分含量反演模型。結(jié)果表明:不同土層的水分含量具有較大差異,棉花冠層對(duì)不同土層水分含量的響應(yīng)程度不同,0~30 cm土層水分含量與棉花冠層含水量的相關(guān)性最強(qiáng),決定系數(shù)達(dá)到0.58;棉花冠層反射率與土壤水含量在可見(jiàn)光波段呈負(fù)相關(guān),近紅外波段呈正相關(guān);在所有以棉花冠層高光譜數(shù)據(jù)的不同變換形式構(gòu)建的不同土層含水量的PLSR反演模型中,以反射率倒數(shù)對(duì)數(shù)所建的模型對(duì)0~30 cm土層和以反射率對(duì)數(shù)所建模型對(duì)0~10 cm土層含水量的預(yù)測(cè)RPD均達(dá)到2.0以上,具有較好的預(yù)測(cè)能力,其余模型的預(yù)測(cè)效果不理想。
[Abstract]:On the basis of water experiment in cotton field, different soil sampling methods were adopted. The hyperspectral data of cotton canopy, soil moisture content data of different depths and cotton canopy water content data were obtained simultaneously, and the relationship between cotton canopy water content and soil water content was analyzed. Based on the correlation between cotton canopy hyperspectral data and soil moisture content, an inversion model of soil moisture content based on cotton canopy hyperspectral data was constructed. The results showed that there were significant differences in soil moisture content among different soil layers. The response degree of cotton canopy to different soil water content was different. The correlation between cotton canopy moisture content and cotton canopy water content was the strongest, and the determining coefficient was 0.58. There was a negative correlation between cotton canopy reflectance and soil water content in visible light band and a positive correlation in near infrared band. In all the PLSR inversion models of different soil moisture content constructed by different transformation form of cotton canopy hyperspectral data. The predicted RPD of 0 ~ 30 cm soil layer and 0 ~ 10 cm soil layer by the model based on the reciprocal logarithm of reflectance is above 2.0. It has better prediction ability, and the other models have not good prediction effect.
【作者單位】： 塔里木大學(xué)經(jīng)濟(jì)與管理學(xué)院資源與環(huán)境經(jīng)濟(jì)研究所;塔里木大學(xué)植物科學(xué)學(xué)院;
【基金】：國(guó)家自然科學(xué)基金項(xiàng)目(41061031,41261083,41361048)資助
【分類號(hào)】：S152.7;TP79
，

本文編號(hào)：1493750