本文選題：遙感圖像　切入點：藍噪聲　出處：《遼寧師范大學》2014年碩士論文　論文類型：學位論文

【摘要】：近年來，隨著高分辨率遙感技術的迅速發(fā)展，高分辨率遙感影像可以針對面實體及其空間位置信息進行較好的觀察。植被是指覆蓋地表的所有植物群落的總稱，與土壤、氣候、地形及水狀況等自然環(huán)境要素密切相關。植被是生態(tài)環(huán)境的最佳反應區(qū)域之一，也是生態(tài)系統(tǒng)的重要組成部分，它在生態(tài)系統(tǒng)中一個非常重要的角色。快速準確地獲取植被信息，先后對于城市建設和綠地規(guī)劃等方面都具有重要意義。但植被結構復雜，種類繁多，具有多尺度集聚以及空間分布的破碎性特征，因此一個好的描述特定尺度像素的森林植被特征的分類方法才可以獲得較為合理的分類結果。 為了實現(xiàn)這一目標，選擇一個合適的尺度來識別遙感圖像植被信息是整個算法的基礎。在開始階段，一種基于藍噪聲特征理論的遙感圖像森林植被紋理的測量方法被用于對遙感圖像森林植被進行尺度提取，與傳統(tǒng)的方法相比減少了算法的時間復雜度。最優(yōu)尺度的提取在整個研究中是非常重要的，因為最優(yōu)的分割比例決定了森林植被的大小和信息提取的精度。要使分割后的多邊形能將原遙感圖像森林植被邊界顯示清楚，既不能太破碎，也不能太模糊。 隨著研究的深入，基于組合各類紋理特征的提取方法的紋理模型被不斷提出，但大多數模型由于方法組合的復雜度高、時間代價過大，而不得不在紋理表達確切度上降低要求。缺少簡潔且確切的結構元尺度來表達模型實現(xiàn)森林植被分割，是目前的主要問題。 本文提出了一種基于藍噪聲理論支持下建立一種快速探測森林植被紋理尺度的方法。經過紋理尺度探測對區(qū)域圖像選擇并縮小持續(xù)探測藍噪聲特征，利用探測到的藍噪聲特征及該尺度確定結構元素的尺度，利用得到的結構元素進行形態(tài)學濾波處理實現(xiàn)遙感圖像森林植被區(qū)域的分割。根據分析得到的森林植被的特定結構元素，然后對遙感圖像進行擊中或擊不中變換處理，使用探測得到的尺度所設計的結構元對森林植被進行分割提取。 進一步，在這些的基礎上可以根據探測出的灌木尺度，，設定特定的結構元來進行灌木分割，這樣在較為復雜的遙感圖像中就可以清晰的提取森林植被與灌木。
[Abstract]:In recent years, with the rapid development of high-resolution remote sensing technology, high-resolution remote sensing images can be used to better observe the surface entities and their spatial location information. Vegetation is one of the best response areas of ecological environment, and it is also an important part of ecosystem. It plays a very important role in the ecosystem. Obtaining vegetation information quickly and accurately is of great significance to urban construction and green space planning, but the vegetation structure is complex and there are many kinds of vegetation. It has the characteristics of multi-scale agglomeration and fragmentation of spatial distribution. Therefore, a good classification method to describe the forest vegetation features with specific size pixels can obtain more reasonable classification results. In order to achieve this goal, selecting a suitable scale to identify the vegetation information of remote sensing images is the basis of the whole algorithm. A remote sensing image forest vegetation texture measurement method based on blue noise feature theory is used to extract forest vegetation from remote sensing image. Compared with the traditional method, the time complexity of the algorithm is reduced. The extraction of the optimal scale is very important in the whole research. Because the optimal segmentation ratio determines the size of forest vegetation and the accuracy of information extraction, the segmented polygon can display the original remote sensing image forest vegetation boundary clearly, neither too broken nor too fuzzy. With the development of the research, the texture model based on the combination of various texture features has been proposed, but most of the models because of the high complexity of the method combination, the time cost is too large. The lack of concise and exact structural element scale to achieve forest vegetation segmentation is the main problem. In this paper, a method based on blue noise theory is proposed to quickly detect the texture scale of forest vegetation. Through texture scale detection, the region image is selected and the blue noise features are reduced. Using the detected blue noise characteristics and the scale of the structural elements, Based on the analysis of the specific structural elements of the forest vegetation, the remote sensing image can be hit or hit. The forest vegetation is segmented and extracted by using the structure element designed by the detected scale. Furthermore, according to the detected shrub scale, a specific structure element can be set to segment the shrub, so that the forest vegetation and shrub can be extracted clearly in the more complex remote sensing image.
【學位授予單位】：遼寧師范大學
【學位級別】：碩士
【學位授予年份】：2014
【分類號】：TP751

【參考文獻】

相關期刊論文 前10條

1 歐陽平;;基于區(qū)域標記的遙感影像植被邊緣檢測算法[J];城市勘測;2011年02期

2 張友靜;樊恒通;;城市植被尺度鑒別與分類研究[J];地理與地理信息科學;2007年06期

3 趙庚星,竇益湘,田文新,張銀輝;衛(wèi)星遙感影像中耕地信息的自動提取方法研究[J];地理科學;2001年03期

4 明冬萍;駱劍承;周成虎;王晶;;高分辨率遙感影像特征分割及算法評價分析[J];地球信息科學;2006年01期

5 楊麗娟,張白樺,葉旭楨;快速傅里葉變換FFT及其應用[J];光電工程;2004年S1期

6 李向榮;王安敏;周桂蓮;;基于快速傅里葉變換的條紋圖像處理研究[J];光學技術;2007年S1期

7 喬鬧生;鄒北驥;鄧磊;曾友兵;鄒劍臣;;一種基于圖像融合的含噪圖像邊緣檢測方法[J];光電子.激光;2012年11期

8 李成;陳仁喜;王秋燕;;改進的基于視覺認知特征的植被識別方法[J];國土資源遙感;2013年02期

9 劉小丹;劉巖;;基于Hough變換和路徑形態(tài)學的城區(qū)道路提取[J];計算機工程;2012年06期

10 李杰;苗長云;武志剛;郭翠娟;;基于數學形態(tài)學的圖像邊緣檢測算法的研究[J];計算機科學;2012年S1期

本文編號：1591464