本文選題：碳收支能力 + 多尺度分割�。� 參考：《廣西師范學(xué)院》2016年碩士論文

【摘要】：全球氣候變暖問題是人類面臨的最艱巨的挑戰(zhàn)之一,通過先進(jìn)的面向?qū)ο蠓诸惙椒▽?duì)中、高分辨率光譜影像進(jìn)行特征信息提取,利用影像信息估算區(qū)域碳收支能力,可以大大提高碳排放與碳匯能力的研究水平,對(duì)于控制區(qū)域氣候變化具有推動(dòng)作用。本文利用面向?qū)ο蠓诸惙椒?以廣西百色市右江區(qū)為研究區(qū)域,選取Landsat 8 OLI和Google Earth影像數(shù)據(jù)提取區(qū)域地物信息,并針對(duì)研究區(qū)地勢(shì)復(fù)雜的特點(diǎn),采用設(shè)置多種尺度參數(shù)的方法,選取最優(yōu)尺度進(jìn)行影像分割。同時(shí),引入隸屬度函數(shù)法、最鄰近分類法和CART決策樹分類器三種方法,基于影像光譜差異、幾何形狀、對(duì)象紋理等特征,逐層逐級(jí)地實(shí)施面向?qū)ο蠓诸惒襟E,隨后加入傳統(tǒng)基于最大似然法的監(jiān)督分類結(jié)果進(jìn)行對(duì)比分析,針對(duì)性的實(shí)現(xiàn)分類結(jié)果精度評(píng)價(jià)分析用以檢驗(yàn)分類結(jié)果的優(yōu)勢(shì)。通過總結(jié)分析前人的地物碳系數(shù)轉(zhuǎn)換關(guān)系并結(jié)合高精度面向?qū)ο蠓诸惤Y(jié)果,構(gòu)建基于土地覆被類型的碳收支能力估算模型,同時(shí)根據(jù)已有的基于CASA模型的碳收支能力估算方法加以精度校驗(yàn),最終估算出右江區(qū)碳收支能力為-399.64萬噸。本文還結(jié)合右江區(qū)行政區(qū)劃、人口分布、DEM等相關(guān)數(shù)據(jù)對(duì)區(qū)域碳收支能力進(jìn)行了專題性剖析。結(jié)果表明:(1)將RS和GIS技術(shù)應(yīng)用與區(qū)域碳收支能力的研究具有很明顯的優(yōu)勢(shì)。多尺度分割和面向?qū)ο蠓诸惙椒芎艽蟪潭壬舷庾V混淆引起的提取誤差,提高中、高分辨率遙感影像的分類運(yùn)行速度,并且解決了傳統(tǒng)分類空間數(shù)據(jù)量大、分類結(jié)果具有“椒鹽現(xiàn)象”等問題,使得不同分類方法可以對(duì)癥下藥,促使碳專題的分類精度有所提升。(2)歸納總結(jié)了國內(nèi)外有關(guān)碳收支能力系數(shù)的研究成果,并將之應(yīng)用于本研究構(gòu)建碳收支估算模型當(dāng)中。利用面向?qū)ο蠓诸惙椒ń庾g出的土地覆被數(shù)據(jù)和碳收支估算模型估算出右江區(qū)碳收支能力結(jié)果。結(jié)果表明林地、草地、耕地等植被覆蓋程度高的地區(qū)擔(dān)負(fù)了研究區(qū)的主要碳匯工作,而建設(shè)用地則因化石燃料的大批量消耗排放了大量的碳源,但右江區(qū)總體上碳匯能力比碳排放能力強(qiáng),有利于區(qū)域生態(tài)系統(tǒng)的穩(wěn)定發(fā)展,較為合理的緩解了區(qū)域氣候變化的威脅。(3)結(jié)合右江區(qū)行政區(qū)劃和DEM數(shù)據(jù)進(jìn)行空間分析。總結(jié)出右江區(qū)碳收支能力總體特征為中心城區(qū)碳收支量大,四周鄉(xiāng)鎮(zhèn)碳收支量小。另一方面,研究分析出影響區(qū)域碳收支能力的主要因素有海拔、坡度坡向和人類活動(dòng)等。具體表現(xiàn)為海拔高、坡度大、人類活動(dòng)量少的地區(qū)碳匯量大、碳排放量小,而海拔低、坡度小、人類活動(dòng)量大的地區(qū)碳排放量大、碳匯量小。綜上所述,面向?qū)ο蠓诸惙椒ㄊ茄芯啃^(qū)域碳收支能力的有效途徑,其分類效果好、速度快,精度高,在區(qū)域碳循環(huán)評(píng)估中具有更好的準(zhǔn)確性和預(yù)見性,可以有效地促進(jìn)碳收支平衡研究領(lǐng)域的發(fā)展。
[Abstract]:Global warming is one of the most difficult challenges facing mankind. Through advanced object-oriented classification methods, feature information is extracted from high-resolution spectral images, and the ability of estimating regional carbon budget is estimated by image information. It can greatly improve the research level of carbon emission and carbon sink ability, and promote the control of regional climate change. In this paper, the object oriented classification method is used to select Landsat 8 OLI and Google Earth image data to extract the regional feature information, taking Youjiang District, Baise City, Guangxi as the research area, and aiming at the complex features of terrain in the study area, the method of setting up various scale parameters is adopted. The optimal scale is selected for image segmentation. At the same time, three methods, membership function method, nearest neighbor classification method and cart decision tree classifier, are introduced. Based on image spectral difference, geometric shape, object texture and other features, the object oriented classification step by step is implemented step by step. Then the traditional supervised classification results based on the maximum likelihood method are compared and analyzed, and the accuracy evaluation analysis of the classification results is carried out to test the advantages of the classification results. Through summing up and analyzing the conversion relation of carbon coefficient of ground objects and combining with the result of high-precision object-oriented classification, a carbon budget estimation model based on land cover type is constructed. At the same time, according to the existing methods of estimating carbon budget and expenditure ability based on CASA model, the accuracy is checked, and finally, the carbon budget capacity of Youjiang region is estimated to be-3.9964 million tons. Based on the relevant data of administrative division and population distribution Dem in Youjiang District, the paper also analyzes the regional carbon budget and expenditure ability. The results show that: (1) the application of RS and GIS technology and the study of regional carbon budget have obvious advantages. Multi-scale segmentation and object-oriented classification method can eliminate the extraction error caused by spectral confusion to a great extent, improve the classification speed of middle and high resolution remote sensing images, and solve the problem of large amount of data in traditional classification space. The classification results have some problems, such as "salt and pepper phenomenon", which make different classification methods fit the case and promote the classification accuracy of carbon topics. (2) summarize the research results of carbon budget capacity coefficient at home and abroad. And it is applied in this study to build carbon budget estimation model. Based on the land cover data and carbon budget estimation model, the capacity of carbon budget in Youjiang region was estimated by using the object-oriented classification method. The results show that the areas with high vegetation coverage, such as woodland, grassland and cultivated land, are responsible for the main carbon sinks in the study area, while the construction land releases a large number of carbon sources due to the massive consumption of fossil fuels. But the ability of carbon sink in Youjiang area is stronger than that of carbon emission, which is beneficial to the stable development of regional ecosystem and reasonably alleviates the threat of regional climate change. (3) the spatial analysis of regional climate change is carried out in combination with administrative division and Dem data of Youjiang district. It is concluded that the overall characteristics of the carbon budget capacity in Youjiang District are that the carbon budget in the central urban area is large, and that in the surrounding towns and villages the carbon budget is small. On the other hand, the main factors influencing regional carbon budget are altitude, slope and human activities. The specific performance is high altitude, large slope, low human activity, large carbon sink, small carbon emission, but low elevation, small slope, large amount of carbon emission and small carbon sink in the area with large human activity. To sum up, the object-oriented classification method is an effective way to study the ability of carbon revenue and expenditure in small regions. It has good classification effect, fast speed and high precision, and has better accuracy and predictability in the assessment of regional carbon cycle. It can effectively promote the development of carbon balance research field.
【學(xué)位授予單位】：廣西師范學(xué)院
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2016
【分類號(hào)】：X87

【相似文獻(xiàn)】

相關(guān)期刊論文 前10條

1 肖慧娟;匡耀求;黃寧生;朱照宇;劉宇;唐金利;;工業(yè)化高速發(fā)展時(shí)期廣州市的碳收支變化初步研究[J];生態(tài)環(huán)境;2006年06期

2 王春權(quán);孟憲民;張曉光;夏丹丹;;陸地生態(tài)系統(tǒng)碳收支/碳平衡研究進(jìn)展[J];資源開發(fā)與市場(chǎng);2009年02期

3 黃耀;中國陸地和近海生態(tài)系統(tǒng)碳收支研究[J];中國科學(xué)院院刊;2002年02期

4 孫曉丹;張為;孔德榮;;發(fā)展碳收支定量認(rèn)證積極應(yīng)對(duì)全球氣候變化[J];法制與社會(huì);2010年10期

5 楊宏玉;涂建軍;楊樂;高紅麗;;土地利用變化對(duì)碳收支平衡的影響研究——以重慶市為例[J];農(nóng)機(jī)化研究;2011年10期

6 付超;于貴瑞;方華軍;王秋鳳;;中國區(qū)域土地利用/覆被變化對(duì)陸地碳收支的影響[J];地理科學(xué)進(jìn)展;2012年01期

7 姚林如;;區(qū)域碳收支結(jié)算的理論和方法探討[J];財(cái)會(huì)研究;2012年13期

8 李春園;楊爽;許艷蘋;;中國近海碳收支、調(diào)控機(jī)理及生態(tài)效應(yīng)研究進(jìn)展[J];中國基礎(chǔ)科學(xué);2012年06期

9 高奇;師學(xué)義;王子凌;李海明;;深圳市碳收支與土地利用變化的協(xié)整分析[J];水土保持研究;2013年06期

10 孫偉;烏日汗;;長三角核心區(qū)碳收支平衡及其空間分異[J];地理研究;2012年12期

相關(guān)會(huì)議論文 前3條

1 周廣勝;王玉輝;王鳳玉;賈丙瑞;許振柱;;不同土地利用方式下的典型草原土壤呼吸作用及其碳收支評(píng)估[A];中國植物學(xué)會(huì)七十周年年會(huì)論文摘要匯編（1933—2003）[C];2003年

2 李芬;夏春海;蔡旭華;賴玉s，

本文編號(hào)：2078825