本文選題：海表溫度 + 劈窗算法��； 參考：《中國(guó)科學(xué)技術(shù)大學(xué)》2017年碩士論文

【摘要】：海洋幅員遼闊,占地面積極大,約是全球面積的70%左右,海洋溫度變化會(huì)在很大程度上影響地球的溫度變化,因此,研究海洋表層溫度至關(guān)重要。海洋蘊(yùn)含著巨大的能量,海水的熱容量非常大,其溫度的微小變化在影響人類生存環(huán)境的同時(shí),會(huì)給全球的天氣帶來(lái)非常大的改變,同時(shí)在一定程度上給局部地區(qū)的天氣帶來(lái)變化。本文分析了海表溫度反演算法的進(jìn)程。在深入分析大量已有成果的同時(shí),結(jié)合MODIS的特點(diǎn),以MODISL1B遙感影像數(shù)據(jù)作為數(shù)據(jù)源,以南海地區(qū)為示范區(qū),對(duì)海表溫度反演的過程作了一系列的研究。具體內(nèi)容如下:(1)以Terra-MODIS數(shù)據(jù)作為數(shù)據(jù)源,基于如今已有的結(jié)果,通過對(duì)比和分析,最終從眾多的反演算法中,選出最適合本文研究區(qū)域的海表溫度反演算法,然后應(yīng)用覆蓋研究區(qū)域的實(shí)測(cè)數(shù)據(jù)做反演實(shí)例結(jié)果的檢驗(yàn)。(2)云是造成紅外反演誤差的主要原因之一,因此,在反演之前,本文做了云的檢測(cè)處理;由于海洋和陸地在紅外波段發(fā)射率差異較大,會(huì)對(duì)反演SST造成影響,因此,本文做了陸地水體分離處理。(3)對(duì)于海表溫度反演軟件的設(shè)計(jì)和實(shí)現(xiàn),以VS 2010為集成環(huán)境,基于C#語(yǔ)言進(jìn)行界面設(shè)計(jì),結(jié)合ArcGIS和ENVI/IDL二次開發(fā)平臺(tái),建立海表溫度反演軟件。在設(shè)計(jì)軟件的過程中,本文給出構(gòu)建海表溫度反演軟件的關(guān)鍵技術(shù)、功能和界面的設(shè)計(jì)方案,并通過調(diào)用GIS組件來(lái)實(shí)現(xiàn)海表溫度反演軟件反演結(jié)果的制圖輸出。(4)對(duì)于反演結(jié)果的驗(yàn)證和分析,以我國(guó)南海海域作為驗(yàn)證區(qū)域范圍來(lái)進(jìn)行。以2014年2月4日標(biāo)準(zhǔn)時(shí)間0250的影像和2014年8月8日標(biāo)準(zhǔn)時(shí)間0245的影像為例,利用軟件進(jìn)行海表溫度計(jì)算,處理結(jié)果表明,此結(jié)果可以真實(shí)地反映試驗(yàn)范圍內(nèi)溫度的分布。用浮標(biāo)測(cè)量值作為驗(yàn)證,結(jié)果顯示,反演結(jié)果與實(shí)測(cè)數(shù)據(jù)的相關(guān)系數(shù)約為0.92,平均誤差MRE為0.69℃,均方根誤差RMSE為0.41℃,最小誤差為0.039℃,最大誤差為1.354℃。表明了本文軟件采用的海表溫度反演算法能夠適用于我國(guó)海域,反演結(jié)果以期對(duì)海洋開發(fā)和探測(cè)提供一定的理論依據(jù)。
[Abstract]:The ocean covers a large area, which is about 70% of the global area. The temperature change of the ocean will affect the temperature change of the earth to a great extent, so it is very important to study the ocean surface temperature.The ocean contains enormous energy, the heat capacity of seawater is very large, the small change of its temperature will affect the living environment of human beings, at the same time, it will bring great changes to the global weather.At the same time, to a certain extent to the local weather changes.The process of sea surface temperature inversion algorithm is analyzed in this paper.In this paper, a series of researches on the inversion of sea surface temperature (SST) are carried out in the South China Sea region and in the South China Sea region, taking the MODISL1B remote sensing image data as the data source and the South China Sea region as the demonstration area, combined with the characteristics of MODIS.The specific contents are as follows: 1) taking Terra-MODIS data as the data source, and based on the existing results, through comparison and analysis, the sea surface temperature inversion algorithm, which is the most suitable for the region studied in this paper, is selected from the numerous inversion algorithms.Then, the cloud is one of the main reasons for the infrared inversion error by using the measured data covering the study area as the test of the inversion examples. Therefore, before inversion, the cloud detection and processing are done in this paper.Because the difference between ocean and land emissivity in infrared band will affect the inversion of SST, this paper designs and implements the software of sea surface temperature inversion, and takes vs 2010 as the integrated environment.Interface design based on C # language, combined with ArcGIS and ENVI/IDL secondary development platform, the establishment of sea surface temperature inversion software.In the process of designing the software, this paper gives the key technology, function and interface design scheme of building the sea surface temperature inversion software.The GIS module is used to realize the mapping output of the inversion result of sea surface temperature inversion software. The verification and analysis of the inversion results are carried out in the South China Sea as the verification area.Taking the images of 0250 standard time on February 4, 2014 and 0245 images of August 8, 2014 as examples, the software is used to calculate the sea surface temperature. The processing results show that the results can truly reflect the temperature distribution in the test range.The results show that the correlation coefficient between the inversion results and the measured data is about 0.92, the average error is 0.69 鈩，

本文編號(hào)：1772108