本文關(guān)鍵詞： 圖像類推 水體 插值法 遙感影像 庫容　出處：《西南交通大學(xué)》2014年碩士論文　論文類型：學(xué)位論文

【摘要】：湖泊不僅是城市水資源的主體,還是農(nóng)業(yè)灌溉的重要水源。然而,由于氣候的變化及人類的過度開發(fā),導(dǎo)致湖泊生態(tài)環(huán)境受到嚴重破壞,水量急劇萎縮。因此,對湖泊水量變化的實時監(jiān)測將會對城市、農(nóng)業(yè)用水起到重要的作用。 Landsat衛(wèi)星遙感影像具有范圍大、周期短、現(xiàn)勢性強、獲取免費等特點,在湖泊水量監(jiān)測方面具有不可比擬的優(yōu)勢。然而,TM影像的空間分辨率僅可達到15m或30m,這就會給水量的實時監(jiān)測帶來很大的誤差。國內(nèi)針對TM影像進行湖泊庫容監(jiān)測方面已經(jīng)有了一定的研究,但如何提高監(jiān)測精度方面還涉及較少。因此,在低空間分辨率影像的前提下提高湖泊水量的監(jiān)測精度將成為研究的關(guān)鍵。 鑒于此,本文開展基于圖像類推法的湖泊庫容提取研究。首先,聯(lián)合圖像類推原理與立方卷積插值法生成超分辨率影像,從而有效提高TM影像的空間分辨率。然后,通過分析超分辨率影像中水體的特性,開展水體信息提取試驗。分析坡度與高差的關(guān)系,通過在湖泊坡度平緩地區(qū)的水位變化來監(jiān)測水量變化,將水位的垂直變化轉(zhuǎn)換為水平方向的變化,并利用區(qū)域內(nèi)插法提取水位線信息。最后,通過結(jié)合由湖泊等深線所生成的DEM模型,有效獲取湖泊的庫容信息。論文以Lake Jackson為實驗對象,利用該方法對多期影像進行試驗,結(jié)果表明,基于圖像類推法的湖泊水位精度平均提高了0.223英尺,庫容精度平均提高了3481.2萬立方英尺。該方法可以在相同空間分辨率影像的條件下提高湖泊水量的監(jiān)測精度,對湖泊庫容量的評估更為精確,將會對水資源分配、利用以及政府的管理和決策起到一定的現(xiàn)實意義。
[Abstract]:Lake is not only the main body of urban water resources, but also an important water source for agricultural irrigation. However, due to climate change and over-exploitation of human beings, the ecological environment of lakes is seriously damaged and the water volume shrinks sharply. Real-time monitoring of lake water change will play an important role in urban and agricultural water use. Landsat satellite remote sensing images have the characteristics of large range, short period, strong reality and free access. However, the spatial resolution of TM image can only reach 15m or 30m, which will bring great error to real-time monitoring of water quantity. There has been some research in the field of measurement. However, how to improve the monitoring accuracy is less concerned. Therefore, improving the monitoring accuracy of lake water quantity under the premise of low spatial resolution image will be the key to the research. In view of this, this paper studies the extraction of lake storage capacity based on image analogy. Firstly, combining the principle of image analogy with cubic convolution interpolation method to generate super-resolution image, thus effectively improve the spatial resolution of TM image. By analyzing the characteristics of water body in super-resolution image, the experiment of extracting water information is carried out, the relationship between slope and elevation difference is analyzed, and the change of water level is monitored through the change of water level in the area with gentle gradient of lake. The vertical change of water level is transformed into the horizontal change, and the information of water level is extracted by region interpolation. Finally, by combining the DEM model generated by the lake isobath, Taking Lake Jackson as the experimental object, the multi-period image is tested by this method. The results show that the accuracy of lake water level based on image analogy method is improved by 0.223 feet on average. The accuracy of reservoir capacity is improved by an average of thirty-four million eight hundred and twelve thousand cubic feet. This method can improve the monitoring accuracy of lake water under the condition of the same spatial resolution image, and the estimation of lake capacity is more accurate, and the allocation of water resources will be improved. Use and government management and decision-making play a certain practical significance.
【學(xué)位授予單位】：西南交通大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2014
【分類號】：P237;P332

【參考文獻】

相關(guān)期刊論文 前10條

1 喬平林,張繼賢,李海濤,曾鈺;水庫容水量遙感監(jiān)測方法研究[J];測繪科學(xué);2003年03期

2 肖艷芳;趙文吉;朱琳;;利用TM影像Band1與Band7提取水體信息[J];測繪科學(xué);2010年05期

3 許榕峰,徐涵秋;ETM+全色波段及其多光譜波段圖像的融合應(yīng)用[J];地球信息科學(xué);2004年01期

4 馬丹;;基于MODIS數(shù)據(jù)的水體提取研究[J];地理空間信息;2008年01期

5 程倩倩;范新南;李慶武;;基于自類推的NSCT域單幅圖像超分辨率重建[J];電子與信息學(xué)報;2011年12期

6 夏雙;阮仁宗;;淡水湖泊濕地水體信息提取方法[J];地理空間信息;2012年05期

7 喻繼業(yè);吳煒;滕奇志;石一興;;基于圖像類推的遙感圖像超分辨率技術(shù)[J];計算機應(yīng)用;2010年01期

8 汪海洋;潘德爐;夏德深;;二維Otsu自適應(yīng)閾值選取算法的快速實現(xiàn)[J];自動化學(xué)報;2007年09期

9 古元亭;吳恩華;;基于圖像類推的超分辨技術(shù)[J];軟件學(xué)報;2008年04期

10 丁志雄;;DEM與遙感相結(jié)合的水庫水位面積曲線測定方法研究[J];水利水電技術(shù);2010年01期

，

本文編號：1549481