北京林业大学学报
北京林業大學學報
북경임업대학학보
JOURNAL OF BEIJING FORESTRY UNIVERSITY
2015年
3期
1-12
,共12页
光学影像%纹理%灰度共生矩阵%最优组合
光學影像%紋理%灰度共生矩陣%最優組閤
광학영상%문리%회도공생구진%최우조합
optical images%texture%gray level co-occurrence matrix%optimal combination
随着光学卫星影像空间分辨率的不断提高,影像纹理特征的重要性日益凸显。然而,纹理是一个非常复杂的空间属性,会随着太阳/观测角度、地形、感兴趣目标及其所处环境的不同而发生显著的变化。此外,不同纹理变量的选择及相应输入参数的设置,如窗口大小、像元间距、方向以及量化等级等都可能在一定程度上决定着影像纹理的利用价值。如何有效利用纹理量及其优化组合是一个值得深入探讨的问题。鉴于此,本文首先全面回顾了影像纹理特征在森林分类、森林结构参数反演以及森林生物量与碳储量的遥感估算等方面的最新研究与应用,并从不同角度肯定了光学影像纹理在林业遥感领域的应用潜力。此外,从纹理变量及其4大输入参数的选择和最优变量组合的判别方面,总结并剖析了当前研究领域中所存在的关键问题,权衡利弊并给出了相应的建议,为相关研究人员将影像纹理信息更有效地应用于林业领域提供参考。
隨著光學衛星影像空間分辨率的不斷提高,影像紋理特徵的重要性日益凸顯。然而,紋理是一箇非常複雜的空間屬性,會隨著太暘/觀測角度、地形、感興趣目標及其所處環境的不同而髮生顯著的變化。此外,不同紋理變量的選擇及相應輸入參數的設置,如窗口大小、像元間距、方嚮以及量化等級等都可能在一定程度上決定著影像紋理的利用價值。如何有效利用紋理量及其優化組閤是一箇值得深入探討的問題。鑒于此,本文首先全麵迴顧瞭影像紋理特徵在森林分類、森林結構參數反縯以及森林生物量與碳儲量的遙感估算等方麵的最新研究與應用,併從不同角度肯定瞭光學影像紋理在林業遙感領域的應用潛力。此外,從紋理變量及其4大輸入參數的選擇和最優變量組閤的判彆方麵,總結併剖析瞭噹前研究領域中所存在的關鍵問題,權衡利弊併給齣瞭相應的建議,為相關研究人員將影像紋理信息更有效地應用于林業領域提供參攷。
수착광학위성영상공간분변솔적불단제고,영상문리특정적중요성일익철현。연이,문리시일개비상복잡적공간속성,회수착태양/관측각도、지형、감흥취목표급기소처배경적불동이발생현저적변화。차외,불동문리변량적선택급상응수입삼수적설치,여창구대소、상원간거、방향이급양화등급등도가능재일정정도상결정착영상문리적이용개치。여하유효이용문리량급기우화조합시일개치득심입탐토적문제。감우차,본문수선전면회고료영상문리특정재삼림분류、삼림결구삼수반연이급삼림생물량여탄저량적요감고산등방면적최신연구여응용,병종불동각도긍정료광학영상문리재임업요감영역적응용잠력。차외,종문리변량급기4대수입삼수적선택화최우변량조합적판별방면,총결병부석료당전연구영역중소존재적관건문제,권형리폐병급출료상응적건의,위상관연구인원장영상문리신식경유효지응용우임업영역제공삼고。
Image texture is becoming more and more important with the increasing spatial resolution of optical satellite images. However, it is a very complex spatial attribute that can vary significantly with solar/viewing geometries, topographic conditions, and the object of interest as well as its location. Besides, the selection of texture variables and the set of their corresponding input parameters, for instance, window sizes, inter-pixel distances, directions and quantization levels, determine the efficiency of image texture. How to apply texture variables and optimize their combination is worth further studying. Therefore, we firstly reviewed the latest researches and applications of image texture in forest classification, inversion of stand structure parameters, and estimation of forest biomass and carbon storage, and prospected the potential of image texture in the remote sensing of forestry from different aspects. In addition, we summarized the critical problems in the current research field based on the selection and optimal combination of texture variables and their input parameters. Some suggestions have also been proposed for further effective applications of image texture in the field of forestry.