【摘要】：城市化的快速發(fā)展,導(dǎo)致城市人口和建筑面積大量增加,自然地表逐漸被大量人工不透水地表例如瀝青路面、混凝土等代替。由于城市地表覆蓋類型的改變,造成地表與空氣中的能量、水分交換發(fā)生了改變,形成了城區(qū)溫度高于郊區(qū)溫度的局地小氣候效應(yīng),即城市熱島效應(yīng)。近些年來,城市熱島效應(yīng)對人類生活環(huán)境造成的影響也是日漸明顯。遙感數(shù)據(jù)能夠獲得大面積城市地面溫度,是一種快捷、有效的技術(shù)手段,本研究中以遙感數(shù)據(jù)來研究喀斯特城市桂林熱環(huán)境變化情況及其影響因子,其目的是為改善桂林人居環(huán)境、科學(xué)的方法進(jìn)行環(huán)境管理提供理論和技術(shù)上的支撐。選取覆蓋桂林市主城區(qū)的美國Landsat 5 TM衛(wèi)星2006年、2009年、2010年三幅圖像,反演地表溫度和描述不透水面、水體、植被的遙感影響參數(shù)。對桂林市地表溫度進(jìn)行正規(guī)化處理,分析地表熱狀況變化情況。對GVI、NDVI、PV、RVI、MSAVI、SAVI、DVI七種植被指數(shù)進(jìn)行剖面分析,通過對其均值和標(biāo)準(zhǔn)差的分析得出植被覆蓋度比較適合作為植被參數(shù)用來與地表溫度進(jìn)行分析,對地域差異并沒有其他植被參數(shù)敏感。定量分析植被覆蓋度與地表溫度的關(guān)系,分別統(tǒng)計不同等級植被覆蓋度區(qū)域的平均溫度,發(fā)現(xiàn)植被覆蓋率比較高的地方,氣溫均值相對較低。對植被變化進(jìn)行時空分析并與地表溫度進(jìn)行回歸分析,發(fā)現(xiàn)植被與地表溫度成負(fù)相關(guān)關(guān)系。定量分析不透水面,發(fā)現(xiàn)不透水面與地表溫度呈明顯的正相關(guān)關(guān)系,利用NDBBI模型提取建筑用地,回歸分析發(fā)現(xiàn)NDBBI與植被水體成負(fù)相關(guān)關(guān)系。定量分析桂林市水體與地表溫度的關(guān)系,發(fā)現(xiàn)水體與地表溫度成明顯的負(fù)相關(guān)關(guān)系。纓帽變換以及主成分分析進(jìn)一步分析與地表溫度相關(guān)的因素,結(jié)果表明綠度分量與地表溫度緊密相關(guān),不透水面對地表溫度的升溫效果超過植被和水體對地表溫度的降溫效果時,城市的熱島效應(yīng)將會更加明顯。鑒于Landsat 5衛(wèi)星熱紅外波段地面分辨率為120m×120m,反演只能獲取該分辨率的地表溫度。為了獲取30m×30m地面分辨率的地表溫度,構(gòu)建120m×120m地表溫度與相關(guān)遙感參數(shù)的神經(jīng)網(wǎng)絡(luò)模型,并將學(xué)習(xí)訓(xùn)練獲得的模型應(yīng)用于輸入30m×30m的遙感參數(shù)。根據(jù)各種地表遙感參數(shù)與地表溫度的相關(guān)系數(shù),以及與地表溫度進(jìn)行回歸擬合的判定系數(shù),選取綠度植被指數(shù)、歸一化植被指數(shù)、修改型調(diào)整植被指數(shù)、比值植被指數(shù)、植被覆蓋度、修改型歸一化差值水體指數(shù)、歸一化差值裸地與建筑用地指數(shù)、不透水面率作為遺傳神經(jīng)網(wǎng)絡(luò)模型進(jìn)行訓(xùn)練和測試的輸入。論文中對選取輸入數(shù)據(jù)的方法進(jìn)行了驗(yàn)證,并且證明以相關(guān)系數(shù)以及回歸分析系數(shù)為判斷原則的方法是可行的。
[Abstract]:With the rapid development of urbanization, the urban population and the building area increase greatly, and the natural surface is gradually replaced by a large number of artificial impervious surfaces such as asphalt pavement, concrete and so on. Because of the change of urban surface cover type, the energy and water exchange between the surface and the air have changed, and the local microclimate effect, which is the urban heat island effect, which is higher than the suburban temperature in urban area, has been formed. In recent years, the urban heat island effect on human living environment is increasingly obvious. Remote sensing data can obtain large area of urban surface temperature, which is a fast and effective technical means. In this study, remote sensing data is used to study the changes of thermal environment and its influencing factors in Guilin, a karst city. The purpose is to provide theoretical and technical support for improving the living environment of Guilin and carrying out scientific environmental management. Three images of Landsat 5 TM satellite covering the main urban area of Guilin in 2006, 2009 and 2010 were selected to retrieve the surface temperature and describe the remote sensing parameters of impermeable surface, water body and vegetation. The surface temperature of Guilin city was regularized and the change of surface heat condition was analyzed. Through the analysis of the mean value and standard deviation of GVI,NDVI,PV,RVI,MSAVI,SAVI,DVI seven vegetation indices, it is concluded that vegetation coverage is more suitable to be used as a vegetation parameter to analyze the surface temperature. No other vegetation parameters are sensitive to regional differences. Quantitative analysis of the relationship between vegetation coverage and surface temperature, statistics of the average temperature of different grades of vegetation coverage areas, it is found that where the vegetation coverage is relatively high, the mean temperature is relatively low. The spatial and temporal analysis of vegetation change and the regression analysis between vegetation and surface temperature showed that vegetation had a negative correlation with surface temperature. Quantitative analysis of impermeable surface showed that impermeable surface was positively correlated with surface temperature. NDBBI model was used to extract construction land. Regression analysis showed that NDBBI had a negative correlation with vegetation water body. Quantitative analysis of the relationship between water body and surface temperature in Guilin City shows that there is an obvious negative correlation between water body and surface temperature. Tasseled hat transformation and principal component analysis (PCA) further analyzed the factors related to the surface temperature. The results showed that the green component was closely related to the surface temperature. The urban heat island effect will be more obvious when the surface temperature warming effect of impermeable water is higher than that of vegetation and water body. Since the ground resolution of the thermal infrared band of Landsat 5 satellite is 120m 脳 120m, the surface temperature can only be obtained by inversion. In order to obtain the surface temperature of 30m 脳 30m ground resolution, the neural network model of 120m 脳 120m surface temperature and related remote sensing parameters is constructed, and the model obtained by learning and training is applied to input 30m 脳 30m remote sensing parameters. According to the correlation coefficient between surface remote sensing parameters and surface temperature, and the decision coefficient of regression fitting with surface temperature, the green vegetation index, normalized vegetation index, modified vegetation index and ratio vegetation index are selected. Vegetation coverage, modified normalized difference water index, normalized difference index of bare land and building land, impermeable surface rate are used as inputs for training and testing of genetic neural network model. In this paper, the method of selecting input data is verified, and it is proved that the method based on correlation coefficient and regression analysis coefficient is feasible.
【學(xué)位授予單位】：廣西師范大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：P423.7;TP183;P407

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 彭華;;基于改進(jìn)的GA優(yōu)化BP神經(jīng)網(wǎng)絡(luò)的煤礦設(shè)備的故障診斷[J];科技通報;2016年11期

2 李召良;段四波;唐伯惠;吳驊;任華忠;閻廣建;唐榮林;冷佩;;熱紅外地表溫度遙感反演方法研究進(jìn)展[J];遙感學(xué)報;2016年05期

3 白蘭東;茍葉培;邵文文;郭云開;伍文;;基于多角度遙感的植被指數(shù)與葉面積指數(shù)的線性關(guān)系研究[J];測繪工程;2016年01期

4 龔珍;胡友健;黎華;;城市水體空間分布與地表溫度之間的關(guān)系研究[J];測繪通報;2015年12期

5 魏寶成;銀山;宋潔;王月紅;;呼和浩特市不同植被指數(shù)與地表溫度的定量遙感關(guān)系[J];水土保持研究;2015年05期

6 李振龍;金雪;王保菊;趙曉華;;基于BP和GA_BP的疲勞駕駛檢測算法對比分析[J];科學(xué)技術(shù)與工程;2015年21期

7 梁保平;馬藝芳;李暉;;桂林市典型園林綠地與水體的降溫效應(yīng)研究[J];生態(tài)環(huán)境學(xué)報;2015年02期

8 周婷;張寅生;高海峰;張騰;馬穎釗;;青藏高原高寒草地植被指數(shù)變化與地表溫度的相互關(guān)系[J];冰川凍土;2015年01期

9 陳利;林輝;;基于K-T變換和主成分變換的植被信息提取[J];中南林業(yè)科技大學(xué)學(xué)報;2014年06期

10 喬治;田光進(jìn);;北京市熱環(huán)境時空分異與區(qū)劃[J];遙感學(xué)報;2014年03期

相關(guān)博士學(xué)位論文 前1條

1 謝啟姣;城市熱島演變及其影響因素研究[D];華中農(nóng)業(yè)大學(xué);2011年

相關(guān)碩士學(xué)位論文 前1條

1 江麗莎;喀斯特城市地表溫度影響因素的遙感反演與分析[D];廣西師范大學(xué);2014年

，

本文編號：2318196