張良培,武 辰
1. 武漢大學(xué)測(cè)繪遙感信息工程國(guó)家重點(diǎn)實(shí)驗(yàn)室,湖北 武漢 430079; 2. 武漢大學(xué)國(guó)際軟件學(xué)院,湖北 武漢 430079
多時(shí)相遙感影像變化檢測(cè)的現(xiàn)狀與展望
張良培1,武 辰2
1. 武漢大學(xué)測(cè)繪遙感信息工程國(guó)家重點(diǎn)實(shí)驗(yàn)室,湖北 武漢 430079; 2. 武漢大學(xué)國(guó)際軟件學(xué)院,湖北 武漢 430079
多時(shí)相遙感影像變化檢測(cè)技術(shù)能夠監(jiān)測(cè)生態(tài)環(huán)境變化、跟蹤城市發(fā)展,對(duì)于研究人類與自然環(huán)境之間的交互關(guān)系有著重要的意義。隨著新型遙感影像的不斷普及,變化檢測(cè)也在高光譜影像變化檢測(cè)和高分辨率影像變化檢測(cè)兩個(gè)方向上有了深入的探索。本文圍繞變化檢測(cè)的基本流程,從預(yù)處理、變化檢測(cè)方法、閾值分割與精度評(píng)價(jià)4個(gè)角度介紹了變化檢測(cè)研究的最新進(jìn)展,并總結(jié)了變化檢測(cè)技術(shù)的主要應(yīng)用領(lǐng)域。最后,對(duì)變化檢測(cè)技術(shù)未來(lái)的發(fā)展進(jìn)行了展望。
變化檢測(cè);遙感影像;多時(shí)相;高分辨率;高光譜
地表生態(tài)系統(tǒng)和人類社會(huì)活動(dòng)都是動(dòng)態(tài)發(fā)展和不斷演變的。實(shí)時(shí)精確地獲取地表變化信息對(duì)于更好地保護(hù)生態(tài)環(huán)境、管理自然資源、研究社會(huì)發(fā)展,以及理解人類活動(dòng)與自然環(huán)境之間的關(guān)系和交互作用有著重要的意義[1-2]。遙感對(duì)地觀測(cè)技術(shù)具有大范圍、長(zhǎng)時(shí)間和周期性監(jiān)測(cè)的能力。因此,利用多時(shí)相遙感數(shù)據(jù)獲取地表地物變化情況的變化檢測(cè)就成為遙感技術(shù)出現(xiàn)最早、應(yīng)用最廣泛的研究領(lǐng)域之一[3-7]。
變化檢測(cè)是通過(guò)對(duì)地物或現(xiàn)象進(jìn)行多次觀測(cè)從而識(shí)別其狀態(tài)變化的過(guò)程[3]。自動(dòng)和半自動(dòng)化的多時(shí)相遙感影像變化檢測(cè)技術(shù)已經(jīng)廣泛應(yīng)用到土地調(diào)查、城市研究、生態(tài)系統(tǒng)監(jiān)測(cè)、災(zāi)害監(jiān)測(cè)評(píng)估以及軍事偵察等應(yīng)用中[8-14]。我國(guó)政府也已經(jīng)高度重視遙感變化檢測(cè)技術(shù)在地理國(guó)情監(jiān)測(cè)中的應(yīng)用,從2010年開(kāi)始國(guó)土資源部每年都會(huì)開(kāi)展全國(guó)土地遙感監(jiān)測(cè)工作,利用多時(shí)相遙感影像變化檢測(cè)技術(shù)持續(xù)更新全國(guó)土地調(diào)查成果[15-16]。
相比于其他遙感數(shù)據(jù)解譯技術(shù),變化檢測(cè)的主要特點(diǎn)是處理和分析不同時(shí)間所獲取的覆蓋同一地區(qū)的多幅遙感影像,其所處理的數(shù)據(jù)量更多(多時(shí)相影像)、數(shù)據(jù)異質(zhì)性更強(qiáng)(成像條件不同所帶來(lái)的數(shù)據(jù)差異)、地物情況更復(fù)雜(變化地物和未變化地物相互混雜)。
變化檢測(cè)基本流程可以大致概括為以下4步。
(1) 預(yù)處理:通過(guò)預(yù)處理步驟進(jìn)行數(shù)據(jù)的配準(zhǔn)和輻射校正,減弱外界成像環(huán)境影響從而簡(jiǎn)化變化檢測(cè)問(wèn)題。
(2) 變化檢測(cè):分析多時(shí)相數(shù)據(jù)中地物的光譜、空間、紋理等特征差異,提取變化強(qiáng)度或“from-to”變化類型等信息。
(3) 閾值分割:將連續(xù)的變化強(qiáng)度利用閾值分割的方式轉(zhuǎn)化為離散的變化信息,生成變化/未變化等語(yǔ)義結(jié)果。
(4) 精度評(píng)價(jià):全面、準(zhǔn)確地評(píng)價(jià)變化檢測(cè)結(jié)果的精度。
國(guó)內(nèi)外學(xué)者已經(jīng)對(duì)變化檢測(cè)問(wèn)題進(jìn)行了大量深入的研究,許多研究工作也模糊了不同步驟之間的邊界,但是目前還未出現(xiàn)一種“萬(wàn)能”的方法來(lái)解決所有問(wèn)題。本文將主要以上述4個(gè)步驟為框架,總結(jié)國(guó)內(nèi)外變化檢測(cè)的相關(guān)研究進(jìn)展,介紹變化檢測(cè)的應(yīng)用領(lǐng)域,并最終對(duì)變化檢測(cè)技術(shù)的發(fā)展進(jìn)行展望。多時(shí)相光學(xué)遙感影像是變化檢測(cè)技術(shù)中使用最廣泛的數(shù)據(jù)源,此外基于多時(shí)相SAR數(shù)據(jù)和LiDAR數(shù)據(jù)的變化檢測(cè)也得到了學(xué)者們的關(guān)注[17-18]。限于篇幅,如果想要將基于SAR和LiDAR等遙感數(shù)據(jù)的變化檢測(cè)研究也進(jìn)行總結(jié),未免會(huì)“掛一漏萬(wàn)”。因此,本文將主要圍繞光學(xué)遙感影像數(shù)據(jù)介紹變化檢測(cè)技術(shù)的最新研究工作。
由于變化檢測(cè)分析的是多時(shí)相遙感影像之間的地物變化和特征相關(guān)性,因此對(duì)于預(yù)處理有其獨(dú)特的要求,主要包括:①配準(zhǔn),保證多時(shí)相影像中同一像素對(duì)應(yīng)同一地理位置地物;②輻射校正,消除不同時(shí)相影像間的輻射差異。預(yù)處理的主要目的是為了減弱外界影響,簡(jiǎn)化變化檢測(cè)問(wèn)題。
配準(zhǔn)誤差是變化檢測(cè)最主要的誤差來(lái)源之一[19],許多研究工作都對(duì)配準(zhǔn)誤差所產(chǎn)生的影響進(jìn)行了深入分析[20]。為了減弱配準(zhǔn)誤差影響,除了影像匹配等專門(mén)的研究方向外,也有許多變化檢測(cè)方法將配準(zhǔn)誤差考慮進(jìn)去,通過(guò)算法建模的方式來(lái)提高變化檢測(cè)精度。文獻(xiàn)[21]利用遙感影像的不同分辨率尺度減弱配準(zhǔn)誤差影響。文獻(xiàn)[22]直接提出基于圖的馬爾科夫場(chǎng)模型,將影像匹配和變化檢測(cè)統(tǒng)一到優(yōu)化過(guò)程中。文獻(xiàn)[23]提出了局部配準(zhǔn)適應(yīng)(local co-registration adjustment,LCRA)策略來(lái)減小配準(zhǔn)誤差影響,這一方法簡(jiǎn)單易用,并可以直接應(yīng)用于大部分高光譜異常變化檢測(cè)算法中。文獻(xiàn)[24]則采用了局部子空間的方式考慮配準(zhǔn)誤差問(wèn)題。
輻射校正是預(yù)處理的另一個(gè)重要步驟[25]。由于多時(shí)相遙感影像獲取時(shí)間不同,包括太陽(yáng)角度、大氣條件等外界成像因素差異會(huì)造成同一地物表現(xiàn)出不同的光譜特征,所造成的“偽變化”是變化檢測(cè)最主要的難題[1,26]?,F(xiàn)有的輻射校正方法可以分為兩類:①將影像DN值轉(zhuǎn)化為地表反射率的絕對(duì)輻射校正;②將目標(biāo)影像輻射值同參考影像輻射值進(jìn)行匹配的相對(duì)輻射校正[27]。絕對(duì)輻射校正需要精確的大氣參數(shù)和復(fù)雜的反演模型,相對(duì)輻射校正則只需要尋找目標(biāo)影像和參考影像間的輻射關(guān)系,更加易于計(jì)算[28]。更有研究指出,相對(duì)輻射校正能夠獲得同絕對(duì)輻射校正相同的處理效果[26,28]。因此,相對(duì)輻射校正得到了更多的關(guān)注和研究。
相對(duì)輻射校正的基本假設(shè)是:未變化地物在多時(shí)相影像同一波段上的輻射值是線性相關(guān)的[26-27,29-30]。但是,真實(shí)變化地物肯定會(huì)對(duì)尋找正確的線性關(guān)系產(chǎn)生影響。因此,如何從影像中自動(dòng)、精確地尋找未變化的校正參考點(diǎn)——偽不變特征點(diǎn)(pseudo-invariant features,PIFs)就是相對(duì)輻射校正研究的關(guān)鍵[25,31]。文獻(xiàn)[32]設(shè)計(jì)了一種自動(dòng)散點(diǎn)圖控制回歸算法選擇未變化像素集。文獻(xiàn)[27]發(fā)展了迭代的PCA方法來(lái)自動(dòng)獲取PIFs。文獻(xiàn)[29]利用影像植被指數(shù)的聚類中心來(lái)計(jì)算線性關(guān)系。文獻(xiàn)[33]深入分析了PIFs同影像場(chǎng)景的輻射特征之間的關(guān)系。多元變化檢測(cè)以及其迭代方法也被用來(lái)提取PIFs作為線性回歸的參考[30,34]。相比于以上提取PIFs的方式,文獻(xiàn)[35]通過(guò)迭代慢特征分析算法獲取每個(gè)像素的未變化概率,并通過(guò)加權(quán)線性回歸的方法利用全部像素直接計(jì)算輻射校正系數(shù),取得了較好的效果。
中低分辨率遙感影像是傳統(tǒng)變化檢測(cè)研究的主要數(shù)據(jù)源,因其數(shù)據(jù)結(jié)構(gòu)簡(jiǎn)單、信息精煉、分辨率適中,可以適用于大多數(shù)變化檢測(cè)任務(wù),是變化檢測(cè)技術(shù)的基礎(chǔ);多時(shí)相高光譜遙感影像能夠提供更加豐富和詳細(xì)的光譜特征及其變化信息,因此能夠?qū)崿F(xiàn)變化類型識(shí)別以及異常變化地物檢測(cè);多時(shí)相高分辨率遙感影像具有空間信息豐富、地物細(xì)節(jié)清晰的特點(diǎn),可以利用空間信息提高變化檢測(cè)精度,也可以檢測(cè)建筑物等特定地物的變化情況。本章將針對(duì)不同數(shù)據(jù)源,總結(jié)相關(guān)變化檢測(cè)研究的最新進(jìn)展。
2.1 中低分辨率影像變化檢測(cè)
針對(duì)中低分辨率遙感影像的變化檢測(cè)方法出現(xiàn)最早、應(yīng)用最廣泛、研究也最為深入[1-3]。按照算法的主要思想,可以總結(jié)為以下4類。
2.1.1 影像代數(shù)法
影像代數(shù)法就是通過(guò)計(jì)算多時(shí)相遙感影像對(duì)應(yīng)波段間的代數(shù)特征來(lái)衡量變化情況。波段差值法是最早出現(xiàn)的變化檢測(cè)方法[3],變化向量分析方法是波段差值法的擴(kuò)展[36],通過(guò)計(jì)算所有波段之間的差值獲得一個(gè)變化特征向量,變化向量的長(zhǎng)度代表變化強(qiáng)度,變化向量的方向代表不同地物變化類型。變化向量分析是目前應(yīng)用最廣泛的方法[37],同時(shí)也為許多其他變化檢測(cè)方法提供基礎(chǔ)數(shù)據(jù)[38]。文獻(xiàn)[39]提出極坐標(biāo)系下區(qū)分不同變化類型的算法框架?;诤朔椒ǖ淖兓蛄糠治鲆驳玫搅藢W(xué)者們的關(guān)注[40-41]。
除了波段差值,波段比值、波段回歸和光譜角也是常用的影像代數(shù)方法[1,42-44]。這些方法能夠提供波段差值以外的變化信息,因此常有研究將其同波段差值結(jié)果進(jìn)行融合[45]。文獻(xiàn)[44]提出了兩種融合光譜角和變化向量分析結(jié)果的策略;文獻(xiàn)[46]用小波方法融合了波段比值結(jié)果和對(duì)數(shù)比值結(jié)果。
2.1.2 影像變換法
相比于原始波段數(shù)據(jù),從多時(shí)相影像數(shù)據(jù)中提取出的特征信息能夠起到突出變化地物、區(qū)分變化類別、提高檢測(cè)精度的效果,這類方法統(tǒng)稱為影像變換法[1-2]。影像變換方法可以從數(shù)據(jù)統(tǒng)計(jì)結(jié)構(gòu)出發(fā),提取出數(shù)據(jù)特征用于變化檢測(cè)。主成分分析方法是較早使用的影像變換方法,根據(jù)特征提取方式不同,可以有:先提取主成分再進(jìn)行差值的PCA差值法[47]、先計(jì)算差值影像再提取主成分的差值PCA法[48]、將多時(shí)相影像疊加到一起進(jìn)行主成分分析的聯(lián)合PCA法[28,47]以及用主成分分析來(lái)尋找兩個(gè)對(duì)應(yīng)波段線性關(guān)系的PCA回歸法[49]等。獨(dú)立成分分析方法也被用于變化檢測(cè)研究中[50]。文獻(xiàn)[51]首先提出了多元變化檢測(cè)方法,并且又進(jìn)一步提出了能夠通過(guò)迭代定權(quán)提高變化檢測(cè)效果的IRMAD方法[52],這一方法已經(jīng)成為目前最有效的非監(jiān)督變化檢測(cè)方法之一[53]。文獻(xiàn)[54]認(rèn)為,通過(guò)最小化未變化地物特征差異能夠有效突出真實(shí)變化信息,并提出了慢特征分析方法(slow feature analysis,SFA),可以有效分離變化和未變化地物,提高檢測(cè)精度[35,55]。
遙感影像的波段具有特定的物理屬性,因此通過(guò)提取具有明確物理意義的特征指數(shù)進(jìn)行變化檢測(cè)也是一種非常有效的方法[56]。文獻(xiàn)[57]采用纓帽變換提取屬性特征進(jìn)行變化檢測(cè)。文獻(xiàn)[58]研究了NDVI指數(shù)在變化檢測(cè)中的效果。文獻(xiàn)[59]對(duì)比了多種植被指數(shù)檢測(cè)土地覆蓋變化的能力。文獻(xiàn)[60]根據(jù)Li-Strahler模型計(jì)算森林結(jié)構(gòu)屬性,分析三峽地區(qū)的植被變化。
2.1.3 分類檢測(cè)法
除了檢測(cè)地物變化的區(qū)域,獲取具體的地物變化類型,即獲取“from-to”變化信息,對(duì)于分析地表變化前后的地物分布情況具有重要的意義,因此監(jiān)督的分類檢測(cè)法在實(shí)際問(wèn)題中得到了非常廣泛的應(yīng)用[10,61]。經(jīng)典的分類檢測(cè)法可以分為兩類:分別獨(dú)立分類再對(duì)比地物類別的分類后變化檢測(cè)[10,38];將多時(shí)相影像疊加到一起進(jìn)行分類的聯(lián)合分類法[62-63]。分類后變化檢測(cè)應(yīng)用較多,但是由于不同時(shí)相影像分類是完全獨(dú)立的,多次分類誤差累計(jì)會(huì)造成變化檢測(cè)精度較低[3];聯(lián)合分類方法將每一種變化類型都看作一類,可以避免誤差累計(jì)問(wèn)題,但變化類型較多會(huì)難以選擇充足的訓(xùn)練樣本,反而無(wú)法在實(shí)際問(wèn)題中得到廣泛應(yīng)用[1,64]。針對(duì)分類檢測(cè)法所存在的問(wèn)題,目前已經(jīng)有了許多相關(guān)的改進(jìn)研究。
文獻(xiàn)[65]統(tǒng)計(jì)得到的地物類型轉(zhuǎn)換概率,并同獨(dú)立分類概率結(jié)合,迭代計(jì)算多時(shí)相地物類別組合的后驗(yàn)概率,提高多時(shí)相影像地物分類精度。文獻(xiàn)[66]采用主動(dòng)學(xué)習(xí)的方法解決多時(shí)相影像分類中樣本選擇的問(wèn)題。文獻(xiàn)[67]將域適應(yīng)理論引入變化檢測(cè)問(wèn)題中。文獻(xiàn)[38]將變化向量分析和分類后變化檢測(cè)相結(jié)合,保持未變化地物的類別一致,在變化地物中采用變化前后的獨(dú)立分類結(jié)果。這一方法簡(jiǎn)單有效,特別在基于多時(shí)相影像的土地利用圖更新問(wèn)題上得到了實(shí)際應(yīng)用[12,38,68]。文獻(xiàn)[69]提出了一種將變化概率同獨(dú)立分類概率相結(jié)合的貝葉斯方法,能夠利用慢特征分析等方法挖掘多時(shí)相影像的相關(guān)性信息,提高變化檢測(cè)及變化類型識(shí)別的精度。
2.1.4 其他方法
以上3類方法是研究最多、應(yīng)用最廣泛的變化檢測(cè)方法。此外,學(xué)者們還在進(jìn)行大量研究和探索。光譜混合分析可以根據(jù)線性混合模型提取中低分辨率影像中地物的亞像素組分分布,也被用于檢測(cè)像素內(nèi)部地物變化情況[70]。文獻(xiàn)[71]用導(dǎo)數(shù)光譜替代原始光譜進(jìn)行變化檢測(cè)。文獻(xiàn)[72]采用信息論中的互信息作為檢測(cè)地物變化的度量。文獻(xiàn)[73]將遺傳算法應(yīng)用到變化檢測(cè)問(wèn)題中。文獻(xiàn)[74]采用影像塊作為地物特征檢測(cè)地物變化。小波分解也被用于實(shí)現(xiàn)多尺度的變化檢測(cè)[75]。
2.2 高光譜影像變化檢測(cè)
高光譜影像包含了豐富且詳細(xì)的光譜特征信息,如何充分利用高光譜影像的優(yōu)勢(shì)實(shí)現(xiàn)多光譜數(shù)據(jù)無(wú)法完成的精細(xì)變化分析以及異常變化檢測(cè),是高光譜變化檢測(cè)的主要科學(xué)問(wèn)題?,F(xiàn)有的高光譜變化檢測(cè)研究可以分為兩類。
2.2.1 異常變化檢測(cè)
異常地物變化區(qū)別于普遍存在的背景地物變化,且數(shù)量稀少。多時(shí)相高光譜遙感影像所包含的精細(xì)光譜特征能夠?yàn)閷?shí)現(xiàn)異常變化檢測(cè)提供可能。文獻(xiàn)[76]將異常變化檢測(cè)流程分為預(yù)測(cè)背景變化的預(yù)測(cè)器和探測(cè)變化殘差異常的探測(cè)器兩部分,并提出了Chronochrome、方差均衡化方法以及基于聚類的改進(jìn)算法。文獻(xiàn)[77]提出一個(gè)異常變化檢測(cè)框架,能夠概括大多數(shù)異常變化檢測(cè)方法,并根據(jù)互信息提出雙曲線算法和其亞像素改進(jìn)算法,這兩種算法是目前最有效的異常變化檢測(cè)方法之一。文獻(xiàn)[78]建立了物理輻射模型,通過(guò)多時(shí)相影像的協(xié)同優(yōu)化檢測(cè)出異常變化,但這種方法使用起來(lái)較為復(fù)雜。文獻(xiàn)[79]用橢球等高分布取代高斯分布進(jìn)行異常變化檢測(cè)算法的推導(dǎo)。文獻(xiàn)[80]提出基于慢特征分析的異常變化檢測(cè)方法,通過(guò)尋找背景地物特征差異最小的投影空間,突出異常變化地物,是目前最有效的異常變化檢測(cè)方法之一。文獻(xiàn)[81]公布了Viareggio數(shù)據(jù)集,其中包含了3幅獲取于同一地區(qū)不同時(shí)間的高光譜影像以及其真實(shí)異常變化地物參考,這也是目前唯一公開(kāi)的高光譜異常變化檢測(cè)數(shù)據(jù)集。
2.2.2 精細(xì)變化分析
通過(guò)挖掘多時(shí)相高光譜影像中豐富的光譜信息能夠精細(xì)地分析地物變化類型,這也是高光譜變化檢測(cè)的重要研究方向。文獻(xiàn)[82]采用多層次聚類的方式來(lái)區(qū)分不同變化類型。文獻(xiàn)[83]將極坐標(biāo)變化向量分析同人工判讀結(jié)合起來(lái),逐層提取變化類型信息。文獻(xiàn)[84]將獨(dú)立成分分析用于提取不同類型地物的變化分布信息。文獻(xiàn)[24]以一個(gè)時(shí)相的影像作為背景,另一個(gè)時(shí)相的影像作為目標(biāo),構(gòu)建多時(shí)相子空間探測(cè)器,并通過(guò)背景子空間不同的構(gòu)建方式來(lái)獲得多樣的變化檢測(cè)效果。
光譜解混是高光譜影像分析的重要理論。文獻(xiàn)[85—87]在一系列工作中將多時(shí)相影像在空間上連接起來(lái),作為一個(gè)樣本數(shù)加倍的數(shù)據(jù)集進(jìn)行光譜解混,并對(duì)比不同時(shí)相同一端元的組分分布情況。文獻(xiàn)[88]將多時(shí)相高光譜影像從光譜維上疊加起來(lái),每種地物變化會(huì)形成一個(gè)新的多時(shí)相端元,再分塊進(jìn)行光譜解混合聚類,得到不同類別的地物變化信息。
2.3 高分辨率影像變化檢測(cè)
高分辨率遙感影像具有地物細(xì)節(jié)清晰、空間信息豐富的特點(diǎn)。因此高分辨率影像變化檢測(cè)研究主要集中于如何充分利用空間信息、保持變化檢測(cè)結(jié)果的完整性,以及檢測(cè)特定地物(建筑物)的變化?,F(xiàn)有研究可以用以下兩類來(lái)概括。
2.3.1 面向?qū)ο蟮淖兓瘷z測(cè)
面向?qū)ο蟮奶幚硎歉叻直媛蔬b感影像解譯領(lǐng)域的重要思想。面向?qū)ο蟮淖兓瘷z測(cè)是以分割地物對(duì)象作為處理單元,綜合考慮對(duì)象的光譜、空間和紋理信息,提高變化檢測(cè)結(jié)果的精度和完整性。文獻(xiàn)[64]系統(tǒng)總結(jié)了面向像素和面向?qū)ο蟮淖兓瘷z測(cè)方法,并對(duì)未來(lái)的發(fā)展進(jìn)行了展望。
面向?qū)ο蟮淖兓瘷z測(cè)中,很重要的一點(diǎn)是如何分割獲得多時(shí)相影像中的地物對(duì)象[89]。文獻(xiàn)[90]將多時(shí)相影像疊加到一起進(jìn)行分割。文獻(xiàn)[91]分別對(duì)影像進(jìn)行分割,再根據(jù)重疊情況獲得最終對(duì)象分割結(jié)果。文獻(xiàn)[92]只對(duì)滑坡發(fā)生后的遙感影像進(jìn)行分割,再進(jìn)行面向?qū)ο蟮淖兓瘷z測(cè)。文獻(xiàn)[93]用已有的土地覆蓋圖作為主題圖層,與最新影像一起進(jìn)行地物分割,并融合地物分類和變化向量分析得到最新的土地利用圖。文獻(xiàn)[94]用不同時(shí)相的全色影像與多光譜影像進(jìn)行交叉融合,再進(jìn)行疊加分割獲得對(duì)象圖。文獻(xiàn)[95]則是將像素級(jí)變化檢測(cè)結(jié)果在對(duì)象內(nèi)部進(jìn)行統(tǒng)計(jì),得到地物對(duì)象的變化情況。文獻(xiàn)[96]對(duì)多時(shí)相影像分別分割,再分別將對(duì)象圖映射到另一時(shí)相影像中,最后融合兩個(gè)方向的變化檢測(cè)結(jié)果。文獻(xiàn)[63]對(duì)比了大量的地物對(duì)象空間特征在變化檢測(cè)中的有效性。
2.3.2 融合空間信息的變化檢測(cè)
除了面向?qū)ο蟮挠跋窠庾g外,通過(guò)加入紋理、空間等特征以及考慮像素空間相關(guān)性的方式,也能夠?qū)⒖臻g信息融入到高分辨率影像變化檢測(cè)中。文獻(xiàn)[97]利用MBI、EVI和NDWI特征提取影像中的建筑物、植被和水體,再通過(guò)計(jì)算影像塊地物統(tǒng)計(jì)直方圖差值的方式,解決多時(shí)相影像觀測(cè)角度不同問(wèn)題,檢測(cè)城市地物變化。文獻(xiàn)[98]用尺度學(xué)習(xí)理論降低變化和未變化地物類內(nèi)方差,增強(qiáng)類間方差,提高變化檢測(cè)精度。文獻(xiàn)[99]提出紋理基元森林的方法來(lái)融合空間上下文信息。文獻(xiàn)[100]采用PCNN神經(jīng)網(wǎng)絡(luò)來(lái)提取空間信息,用于變化檢測(cè)。文獻(xiàn)[101]用PCNN神經(jīng)網(wǎng)絡(luò)來(lái)對(duì)影像分別進(jìn)行二值化,再采用歸一化轉(zhuǎn)動(dòng)慣量特征來(lái)進(jìn)行變化檢測(cè)。文獻(xiàn)[102]使用條件隨機(jī)場(chǎng)挖掘影像多種光譜/空間特征,在減少由于影像分辨率提升導(dǎo)致的虛警點(diǎn)問(wèn)題的同時(shí),保持變化地物的細(xì)節(jié)輪廓信息。
由于高分辨率影像中可以區(qū)分出獨(dú)立的地物,特別是城市的主要地物——建筑物,因此如何提取多時(shí)相高分辨率影像中的空間形狀特征,對(duì)城市建筑物的變化進(jìn)行檢測(cè),也是變化檢測(cè)研究中的重要方向。文獻(xiàn)[103]提取出建筑物的形狀,提出容錯(cuò)性變化檢測(cè)方法。文獻(xiàn)[104]專門(mén)針對(duì)建筑物變化檢測(cè)提出了建筑物變化指數(shù)。文獻(xiàn)[105]采用基于信息論的方法,對(duì)建筑物指數(shù)特征影像進(jìn)行變化檢測(cè)。文獻(xiàn)[106]將慢特征分析結(jié)果同MBI建筑物指數(shù)進(jìn)行融合,得到變化的建筑物區(qū)域。文獻(xiàn)[107]結(jié)合遙感影像所提取的建筑物輪廓特征和GIS數(shù)據(jù)檢測(cè)建筑物是否變化,主要應(yīng)用于地震災(zāi)后的建筑物監(jiān)測(cè)。
閾值分割是指根據(jù)變化檢測(cè)方法所獲得的變化強(qiáng)度信息劃分變化/未變化地物的過(guò)程。許多變化檢測(cè)研究以單波段影像差值或變化向量分析強(qiáng)度圖作為輸入,以變化強(qiáng)度分割作為主要研究?jī)?nèi)容,也可以認(rèn)為屬于閾值分割研究[108]。在變化檢測(cè)研究中,二類k均值算法和大津閾值算法是兩種簡(jiǎn)單且有效的自動(dòng)閾值分割算法[54-55,74]。文獻(xiàn)[109]假設(shè)變化強(qiáng)度圖是由兩個(gè)高斯分布混合而成,提出了經(jīng)典的EM閾值分割方法。文獻(xiàn)[110]使用馬爾科夫隨機(jī)場(chǎng)來(lái)融合多個(gè)自動(dòng)閾值分割的結(jié)果。文獻(xiàn)[111]引入水平集理論對(duì)變化強(qiáng)度圖進(jìn)行分割。文獻(xiàn)[112]進(jìn)一步將基于EM算法的水平集用于變化檢測(cè)中。
在變化檢測(cè)的最后,需要對(duì)其結(jié)果進(jìn)行精度評(píng)價(jià),分析變化檢測(cè)結(jié)果的可靠性。目前常用的精度評(píng)價(jià)方法包含以下3種:
(1) 混淆矩陣:混淆矩陣是分類精度評(píng)價(jià)中應(yīng)用最廣泛的方法。在變化檢測(cè)中,可以將變化/未變化二值結(jié)果看作是一個(gè)二類分類結(jié)果,使用混淆矩陣、總體精度和Kappa系數(shù)來(lái)評(píng)價(jià)變化檢測(cè)精度[38,113]。在“from-to”變化類型分析問(wèn)題中,可以將每一個(gè)“from-to”變化類型看作一個(gè)類別,同樣使用混淆矩陣進(jìn)行分析[68]。
(2) ROC曲線:ROC曲線是一種不受分割閾值影響,能夠評(píng)價(jià)區(qū)分變化能力強(qiáng)弱的方法[76-77]。通過(guò)遍歷閾值,獲取每一個(gè)“檢測(cè)率-誤檢率”數(shù)據(jù)點(diǎn),畫(huà)出一條ROC曲線。越靠近左上角的曲線代表著在同一誤檢率下,檢測(cè)率更高,即變化檢測(cè)能力更強(qiáng)。同時(shí),通過(guò)計(jì)算曲線下面積(area under curve,AUC),可以定量評(píng)價(jià)變化檢測(cè)能力的強(qiáng)弱[54]。
(3) 檢測(cè)率:通過(guò)成功檢測(cè)出來(lái)的變化地物比例來(lái)評(píng)價(jià)算法的效果,用于評(píng)價(jià)二值變化檢測(cè)結(jié)果,一般同誤檢率搭配使用[106]。為了綜合兩個(gè)指標(biāo),也可以采用F-score來(lái)進(jìn)行總體的精度評(píng)價(jià)[69]。
變化檢測(cè)是遙感對(duì)地觀測(cè)中出現(xiàn)最廣、應(yīng)用最廣泛的技術(shù)之一,在環(huán)境、資源、城市、災(zāi)害、軍事等領(lǐng)域都起到了重要的作用??偨Y(jié)起來(lái),遙感變化檢測(cè)技術(shù)具有表1所示應(yīng)用領(lǐng)域。
表1遙感變化檢測(cè)技術(shù)的實(shí)際應(yīng)用領(lǐng)域
Tab.1Practicalapplicationofremotesensingchangedetectiontechnology
應(yīng)用領(lǐng)域具體方向應(yīng)用實(shí)例土地調(diào)查土地覆蓋/土地利用監(jiān)測(cè)[8,10,38,47,114—115]城市研究城市擴(kuò)展和不透水面變化檢測(cè)[12,61,107,116]建筑物變化檢測(cè)[105,107,117]生態(tài)系統(tǒng)監(jiān)測(cè)森林覆蓋變化檢測(cè)[60,90,118]湖泊環(huán)境監(jiān)測(cè)[14,119]濕地環(huán)境監(jiān)測(cè)[120—121]海岸環(huán)境監(jiān)測(cè)[122]旱地變化監(jiān)測(cè)[123]植被變化監(jiān)測(cè)[58,124]沙地變化監(jiān)測(cè)[125]自然保護(hù)區(qū)監(jiān)測(cè)[126]冰層覆蓋情況監(jiān)測(cè)[127]災(zāi)害檢測(cè)和評(píng)估山火檢測(cè)和影響評(píng)估[11,128]滑坡檢測(cè)[13,92]地震損害評(píng)估[129]海嘯損害評(píng)估[130]漏油區(qū)域監(jiān)測(cè)和影響評(píng)估[131]軍事應(yīng)用戰(zhàn)爭(zhēng)對(duì)環(huán)境的影響評(píng)估[132]核試驗(yàn)場(chǎng)檢測(cè)[9]軍事打擊效果評(píng)估[133]
結(jié)合當(dāng)前最新的計(jì)算機(jī)理論和技術(shù),變化檢測(cè)技術(shù)在以下幾個(gè)方面還能夠煥發(fā)出新的活力。
6.1 場(chǎng)景變化檢測(cè)
遙感影像場(chǎng)景分類是指根據(jù)影像內(nèi)部地物的空間和結(jié)構(gòu)分布模式來(lái)識(shí)別場(chǎng)景的語(yǔ)義類別,即工業(yè)區(qū)、商業(yè)區(qū)等土地利用類型[134]。遙感場(chǎng)景分類目前已經(jīng)成為遙感數(shù)據(jù)解譯技術(shù)新的熱點(diǎn)[135]。但是,目前還少有學(xué)者對(duì)多時(shí)相遙感影像的場(chǎng)景變化檢測(cè)技術(shù)進(jìn)行研究。場(chǎng)景變化檢測(cè)能夠在語(yǔ)義層次檢測(cè)區(qū)域土地利用類型的改變情況,例如在城市發(fā)展中所出現(xiàn)的棚戶區(qū)改造、工業(yè)區(qū)外遷、商業(yè)區(qū)改造等,對(duì)于監(jiān)測(cè)城市發(fā)展和輔助科學(xué)規(guī)劃有著重要的意義[136]。
在場(chǎng)景變化檢測(cè)中,影像場(chǎng)景內(nèi)部地物的變化不會(huì)直接造成場(chǎng)景類別的變化,例如一棟建筑物的修建不會(huì)令居民區(qū)變成工業(yè)區(qū)。因此,必須從語(yǔ)義層次上研究變化檢測(cè)算法。文獻(xiàn)[136]提出了基于BOVW模型的場(chǎng)景變化檢測(cè)方法,并對(duì)比分析了不同字典構(gòu)建方法對(duì)最終結(jié)果的影響。為了解決獨(dú)立分類所帶來(lái)的誤差累計(jì)問(wèn)題,文獻(xiàn)[55]提出核化慢特征分析方法,并通過(guò)貝葉斯理論融合場(chǎng)景變化概率和場(chǎng)景分類概率。除了目前已有的少數(shù)工作外,場(chǎng)景變化檢測(cè)還需要在聯(lián)合字典編碼、高維復(fù)雜場(chǎng)景特征變化檢測(cè)等方面進(jìn)行研究。
6.2 高光譜變化檢測(cè)
隨著高光譜傳感器的逐漸普及,多時(shí)相高光譜影像的獲取將更加容易,因此高光譜變化檢測(cè)研究將會(huì)迎來(lái)發(fā)展的機(jī)遇。高光譜變化檢測(cè)研究的主要突破方向是,利用豐富的光譜信息檢測(cè)像素內(nèi)部地物組分的變化情況,以及非監(jiān)督的地物變化類型分析。線性光譜混合模型是高光譜數(shù)據(jù)解譯的重要理論,如何將光譜解混同變化檢測(cè)結(jié)合起來(lái),將會(huì)成為高光譜變化檢測(cè)的研究重點(diǎn)。雖然目前也有一些相關(guān)的工作[85-88],但還沒(méi)有建立起多時(shí)相影像中端元的光譜相關(guān)性,也沒(méi)有發(fā)展出多時(shí)相光譜混合模型,所以理論上還需要進(jìn)一步的創(chuàng)新。
6.3 分類變化檢測(cè)方法的改進(jìn)
由于可以提供詳細(xì)的“from-to”變化信息,分類變化檢測(cè)法在遙感地學(xué)研究中得到了廣泛的應(yīng)用[10,61]。分類后變化檢測(cè)雖然存在誤差累計(jì)造成的變化檢測(cè)精度不高問(wèn)題,但是由于原理簡(jiǎn)單、易于理解,依然是目前使用最多的方法。多時(shí)相影像間的相關(guān)性信息能夠有效提高分類后變化檢測(cè)的精度,光譜差異較小的地物更可能屬于同一類別,光譜差異較大的地物更大概率發(fā)生類別變化[38,55,68]。因此,如何充分挖掘多時(shí)相影像的相關(guān)性信息,發(fā)展簡(jiǎn)單、有效、穩(wěn)健的分類后變化檢測(cè)改進(jìn)算法,是非常具有實(shí)用價(jià)值的研究方向。
此外,覆蓋同一區(qū)域的時(shí)間序列影像數(shù)據(jù)集也越來(lái)越容易獲取。除了時(shí)間序列分析方法外,也需要研究基于分類的時(shí)間序列影像變化檢測(cè)方法,利用時(shí)間序列影像的時(shí)空相關(guān)信息提高地物分類的精度和連續(xù)性[137]。
6.4 多源多分辨率變化檢測(cè)
大多數(shù)變化檢測(cè)研究都使用同一傳感器的多時(shí)相影像數(shù)據(jù)。但是,同源遙感影像可能由于觀測(cè)難度、成本、覆蓋周期等原因,無(wú)法獲取合適的重復(fù)觀測(cè)數(shù)據(jù)。因此,研究多源多分辨率變化檢測(cè)能夠大大擴(kuò)展變化檢測(cè)技術(shù)的應(yīng)用范圍。
多源多時(shí)相數(shù)據(jù)由于成像機(jī)理、觀測(cè)特征的不同,一方面可以提供多角度的觀測(cè)信息,另一方面卻又很難建立起地物特征的相關(guān)性。文獻(xiàn)[113]通過(guò)兩個(gè)對(duì)應(yīng)的稀疏表達(dá)字典建立特征相關(guān)性。文獻(xiàn)[40]通過(guò)定義多源數(shù)據(jù)的核函數(shù)實(shí)現(xiàn)變化檢測(cè)。更多和更加通用的多源變化檢測(cè)研究有望出現(xiàn)在高光譜-高分辨率影像、SAR-光學(xué)影像以及LiDAR-光學(xué)影像等多時(shí)相數(shù)據(jù)分析中。
分辨率較高的數(shù)據(jù)往往無(wú)法像低分辨率遙感影像一樣具有較短的重訪周期,發(fā)展多分辨率變化檢測(cè)能夠提高多時(shí)相數(shù)據(jù)的觀測(cè)密度,并獲得高分辨率的變化檢測(cè)結(jié)果。文獻(xiàn)[138]將高分辨率影像的分類結(jié)果同低分辨率影像的解混結(jié)果進(jìn)行結(jié)合,分析低分辨率影像的亞像素地物變化。文獻(xiàn)[139]利用高分辨率影像的硬分類結(jié)果和低分辨率影像的軟分類結(jié)果進(jìn)行亞像素制圖,再對(duì)比亞像素分類圖獲得變化檢測(cè)結(jié)果?,F(xiàn)有研究工作主要還以監(jiān)督變化檢測(cè)方法為主,如何非監(jiān)督地融合多分辨率影像,實(shí)現(xiàn)亞像素級(jí)變化檢測(cè),還有待進(jìn)一步探索。
6.5 深度學(xué)習(xí)變化檢測(cè)
深度學(xué)習(xí)能夠自動(dòng)、多層次地提取復(fù)雜對(duì)象的抽象特征,大幅度提高模式識(shí)別精度。深度學(xué)習(xí)理論也能夠應(yīng)用到變化檢測(cè)領(lǐng)域,從多時(shí)相影像中提取空間-光譜的一體化特征,以及建立多時(shí)相地物特征的非線性相關(guān)性。文獻(xiàn)[43]利用深度神經(jīng)網(wǎng)絡(luò)從多時(shí)相影像所提取的影像塊中學(xué)習(xí)特征,再利用多時(shí)相地物特征的光譜角和極坐標(biāo)方向來(lái)識(shí)別是否發(fā)生變化、區(qū)分不同變化類型。文獻(xiàn)[140]利用2層SAE神經(jīng)網(wǎng)絡(luò)學(xué)習(xí)多源數(shù)據(jù)間的特征變換模型,建立起多源遙感影像特征的相關(guān)性。文獻(xiàn)[141]提出了一個(gè)包含卷積層和耦合層的神經(jīng)網(wǎng)絡(luò),通過(guò)卷積層來(lái)提取特征,耦合層來(lái)學(xué)習(xí)未變化地物的一致性特征,最后計(jì)算多時(shí)相數(shù)據(jù)經(jīng)過(guò)神經(jīng)網(wǎng)絡(luò)變換后的特征差異,提取變化地物。
雖然已有工作將深度學(xué)習(xí)用來(lái)學(xué)習(xí)多時(shí)相數(shù)據(jù)間的相關(guān)性特征,但是深度學(xué)習(xí)變化檢測(cè)研究還有非常巨大的潛力。在高分辨率影像變化檢測(cè)中,應(yīng)該更多利用多層次的地物空間/形狀特征,因此需要研究更加深層次的網(wǎng)絡(luò)結(jié)構(gòu)。此外,也可以更加深入地發(fā)揮深度學(xué)習(xí)神經(jīng)網(wǎng)絡(luò)在語(yǔ)義理解方面的能力,利用遙感大數(shù)據(jù)學(xué)習(xí)強(qiáng)大和通用的場(chǎng)景變化檢測(cè)模型。
變化檢測(cè)作為最早出現(xiàn)、也是應(yīng)用最廣泛的遙感技術(shù)之一,一直以來(lái)都是理論算法研究和地學(xué)分析領(lǐng)域的熱門(mén)話題。隨著新型遙感影像的不斷普及,變化檢測(cè)也在高光譜影像變化檢測(cè)和高分辨率影像變化檢測(cè)兩個(gè)方向上有了深入的探索。本文圍繞著變化檢測(cè)的基本流程,從預(yù)處理、變化檢測(cè)方法、閾值分割與精度評(píng)價(jià)4個(gè)角度介紹了變化檢測(cè)最新的研究進(jìn)展。特別是針對(duì)不同的遙感數(shù)據(jù)類型,本文詳細(xì)總結(jié)了中低分辨率影像變化檢測(cè)方法、高光譜影像變化檢測(cè)方法和高分辨率影像變化檢測(cè)的相關(guān)研究工作。其中,中低分辨率遙感影像變化檢測(cè)是變化檢測(cè)技術(shù)的基礎(chǔ),高光譜影像變化檢測(cè)和高分辨率影像變化檢測(cè)都是結(jié)合影像的優(yōu)勢(shì)和特點(diǎn)所進(jìn)行的改進(jìn)和發(fā)展。此外,本文還總結(jié)了變化檢測(cè)的應(yīng)用領(lǐng)域及其具體方向。
最后,本文對(duì)變化檢測(cè)技術(shù)的未來(lái)發(fā)展進(jìn)行了展望。筆者認(rèn)為,變化檢測(cè)技術(shù)在以下5個(gè)方面還具有很大的潛力:①場(chǎng)景變化檢測(cè),在語(yǔ)義層次對(duì)土地利用變化情況進(jìn)行檢測(cè)和分析;②高光譜變化檢測(cè),結(jié)合光譜混合模型實(shí)現(xiàn)非監(jiān)督的變化類型分析與亞像素變化檢測(cè);③分類變化檢測(cè)方法的改進(jìn),充分利用多時(shí)相影像間的時(shí)空相關(guān)性提高多時(shí)相分類結(jié)果的一致性和“from-to”變化信息檢測(cè)精度;④多源多分辨率變化檢測(cè),研究通用的變化檢測(cè)理論與方法,充分利用不同觀測(cè)機(jī)理和不同分辨率的多時(shí)相遙感數(shù)據(jù);⑤深度學(xué)習(xí)變化檢測(cè),用深度神經(jīng)網(wǎng)絡(luò)提取多時(shí)相影像的光譜/空間一致性特征,獲得高精度的變化檢測(cè)結(jié)果。
[1] LU D,MAUSEL P,BRONDZIO E,et al. Change Detection Techniques[J]. International Journal of Remote Sensing, 2004, 25(12): 2365-2401.
[2] COPPIN P,JONCKHEERE I,NACKAERTS K,et al. Digital Change Detection Methods in Ecosystem Monitoring: A Review[J]. International Journal of Remote Sensing, 2004, 25(9): 1565-1596.
[3] SINGH A. Digital Change Detection Techniques Using Remotely-sensed Data[J]. International Journal of Remote Sensing, 1989, 10(6): 989-1003.
[4] CRAPPER PF,HYNSON KC. Change Detection Using Landsat Photographic Imagery[J]. Remote Sensing of Environment, 1983, 13(4): 291-300.
[5] 李德仁. 利用遙感影像進(jìn)行變化檢測(cè)[J]. 武漢大學(xué)學(xué)報(bào)(信息科學(xué)版), 2003, 28(S1): 7-12.
LI Deren. Change Detection from Remote Sensing Images[J]. Geomatics and Information Science of Wuhan University, 2003, 28(S1): 7-12.
[6] KENNEDY R E,TOWNSEND P A,GROSS J E,et al. Remote Sensing Change Detection Tools for Natural Resource Managers: Understanding Concepts and Tradeoffs in the Design of Landscape Monitoring Projects[J]. Remote Sensing of Environment, 2009, 113(7): 1382-1396.
[7] 周啟鳴. 多時(shí)相遙感影像變化檢測(cè)綜述[J]. 地理信息世界, 2011, 9(2): 28-33.
ZHOU Qiming. Review on Change Detection Using Multi-temporal Remotely Sensed Imagery[J]. Geomatics World, 2011, 9(2): 28-33.
[8] LI Haixing,XIAO Pengfeng,FENG Xuezhi,et al.Using Land Long-term Data Records to Map Land Cover Changes in China over 1981—2010[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2017, 10(4): 1372-1389.
[9] ZELINSKI ME,HENDERSON J,SMITH M. Use of Landsat 5 for Change Detection at 1998 Indian and Pakistani Nuclear Test Sites[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2014, 7(8): 3453-3460.
[10] HU Jinrong,ZHANG Yuanzhi. Seasonal Change of Land-use/Land-cover (LULC) Detection Using MODIS Data in Rapid Urbanization Regions: A Case Study of the Pearl River Delta Region (China)[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2013, 6(4): 1913-1920.
[11] VOGELMANN J E,KOST J R,TOLK B, et al. Monitoring Landscape Change for LANDFIRE Using Multi-temporal Satellite Imagery and Ancillary Data[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2011, 4(2): 252-264.
[12] XIAN G, HOMER C. Updating the 2001 National Land Cover Database Impervious Surface Products to 2006 Using Landsat Imagery Change Detection Methods[J]. Remote Sensing of Environment, 2010, 114(8): 1676-1686.
[13] LI Zhongbin, SHI Wenzhong, MYINT S W,et al. Semi-automated Landslide Inventory Mapping from Bitemporal Aerial Photographs Using Change Detection and Level Set Method[J]. Remote Sensing of Environment, 2016(175): 215-230.
[14] SONG Chunqiao, HUANG Bo, KE Linghong, et al. Remote Sensing of Alpine Lake Water Environment Changes on the Tibetan Plateau and Surroundings: A Review[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2014(92): 26-37.
[15] 國(guó)土資源部地籍管理司.國(guó)土資源部關(guān)于開(kāi)展2010年度全國(guó)土地變更調(diào)查與遙感監(jiān)測(cè)工作的通知[M]. 北京: 國(guó)土資源部, 2010.
Ministry of Land and Resources of the People’s Republic of China.The Notice for National Land-use Change Survey and Remote Sensing Monitoring from Ministry of Land and Resources in 2010[M].Beijing:Ministry of Land and Resources,2010.
[16] 國(guó)土資源部地籍管理司.國(guó)土資源部關(guān)于開(kāi)展2016年度全國(guó)土地變更調(diào)查與遙感監(jiān)測(cè)工作的通知[M]. 北京: 國(guó)土資源部, 2016.
Ministry of Land and Resources of the People’s Republic of China.The Notice for National Land-use Change Survey and Remote Sensing Monitoring from Ministry of Land and Resources in 2016[M].Beijing: Ministry of Land and Resources,2016.
[17] BAZI Y,BRUZZONE L,MELGANI F. An Unsupervised Approach Based on the Generalized Gaussian Model to Automatic Change Detection in Multitemporal SAR Images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2005, 43(4): 874-887.
[18] VUT T, MATSUOKA M,YAMAZAKI F. LiDAR-based Change Detection of Buildings in Dense Urban Areas[C]∥Proceedings of 2004 IEEE International Geoscience and Remote Sensing Symposium. Anchorage:IEEE,2004:3413-3416.
[19] DAI Xiaolong, KHORRAM S. The Effects of Image Misregistration on the Accuracy of Remotely Sensed Change Detection[J]. IEEE Transactions on Geoscience and Remote Sensing, 1998, 36(5): 1566-1577.
[20] CHEN Gang, ZHAO Kaiguang, POWERS R. Assessment of the Image Misregistration Effects on Object-based Change Detection[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2014(87): 19-27.
[21] MARCHESI S,BOVOLO F,BRUZZONE L. A Context-Sensitive Technique Robust to Registration Noise for Change Detection in VHR Multispectral Images[J]. IEEE Transactions on Image Processing, 2010, 19(7): 1877-1889.
[22] VAKALOPOULOU M,KARANTZALOS K,KOMODAKIS N,et al. Graph-based Registration, Change Detection, and Classification in very High Resolution Multitemporal Remote Sensing Data[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2016, 9(7): 2940-2951.
[23] THEILER J, WOHLBERG B. Local Coregistration Adjustment for Anomalous Change Detection[J]. IEEE Transactions on Geoscience and Remote Sensing, 2012, 50(8): 3107-3116.
[24] WU Chen, DU Bo, ZHANG Liangpei. A Subspace-based Change Detection Method for Hyperspectral Images[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2013, 6(2): 815-830.
[25] YUAN Ding, ELVIDGE C D. Comparison of Relative Radiometric Normalization Techniques[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 1996, 51(3): 117-126.
[26] SONG Conghe, WOODCOCK C E,SETO K C,et al. Classification and Change Detection Using Landsat TM Data: When and How to Correct Atmospheric Effects?[J]. Remote Sensing of Environment, 2001, 75(2): 230-244.
[27] DU Yong, TEILLET P M,CIHLAR J. Radiometric Normalization of Multitemporal High-resolution Satellite Images with Quality Control for Land Cover Change Detection[J]. Remote Sensing of Environment, 2002, 82(1): 123-134.
[28] COLLINS J B,WOODCOCK C E. An Assessment of Several Linear Change Detection Techniques for Mapping Forest Mortality Using Multitemporal Landsat TM Data[J]. Remote Sensing of Environment, 1996, 56(1): 66-77.
[29] CHEN Xuexia, VIERLING L, DEERING D. A Simple and Effective Radiometric Correction Method to Improve Landscape Change Detection across Sensors and across Time[J]. Remote Sensing of Environment, 2005, 98(1): 63-79.
[30] CANTY M J,NIELSEN A A,SCHMIDT M. Automatic Radiometric Normalization of Multitemporal Satellite Imagery[J]. Remote Sensing of Environment, 2004, 91(3-4): 441-451.
[31] SCHOTT J R,SALVAGGIO C, VOLCHOK W J. Radiometric Scene Normalization Using Pseudoinvariant Features[J]. Remote Sensing of Environment, 1988, 26(1): 1-16.
[32] ELVIDGE C D,YUAN D,WEERACKOON R D,et al. Relative Radiometric Normalization of Landsat Multispectral Scanner (MSS) Data Using an Automatic Scattergram-Controlled Regression[J]. Photogrammetric Engineering and Remote Sensing, 1995, 61(10): 1255-1260.
[33] PHILPOT W,ANSTY T. Analytical Description of Pseudoinvariant Features[J]. IEEE Transactions on Geoscience and Remote Sensing, 2013, 51(4): 2016-2021.
[34] CANTY M J,NIELSEN A A. Automatic Radiometric Normalization of Multitemporal Satellite Imagery with the Iteratively Re-weighted MAD Transformation[J]. Remote Sensing of Environment, 2008, 112(3): 1025-1036.
[35] ZHANG Liangpei, WU Chen, DU Bo. Automatic Radiometric Normalization for Multitemporal Remote Sensing Imagery with Iterative Slow Feature Analysis[J]. IEEE Transactions on Geoscience and Remote Sensing, 2014, 52(10): 6141-6155.
[36] MALILA W A. Change Vector Analysis:An Approach for Detecting Forest Changes with Landsat[C]∥LARS Symposia. West Lafayette:Purdue University,1980:326-336.
[37] BOVOLO F, BRUZZONE L. A Theoretical Framework for Unsupervised Change Detection Based on Change Vector Analysis in the Polar Domain[J]. IEEE Transactions on Geoscience and Remote Sensing, 2007, 45(1): 218-236.
[38] XIAN G, HOMER C, FRY J. Updating the 2001 National Land Cover Database Land Cover Classification to 2006 by Using Landsat Imagery Change Detection Methods[J]. Remote Sensing of Environment, 2009, 113(6): 1133-1147.
[39] BOVOLO F, MARCHESI S, BRUZZONE L. A Framework for Automatic and Unsupervised Detection of Multiple Changes in Multitemporal Images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2012, 50(6): 2196-2212.
[40] CAMPS-VALLS G, GOMEZ-CHOVA L, MUNOZ-MARI J, et al. Kernel-based Framework for Multitemporal and Multisource Remote Sensing Data Classification and Change Detection[J]. IEEE Transactions on Geoscience and Remote Sensing, 2008, 46(6): 1822-1835.
[41] VOLPI M,TUIA D,CAMPS-VALLS G,et al. Unsupervised Change Detection with Kernels[J]. IEEE Geoscience and Remote Sensing Letters, 2012, 9(6): 1026-1030.
[42] RIDD M K,LIU Jiajun. A Comparison of Four Algorithms for Change Detection in an Urban Environment[J]. Remote Sensing of Environment, 1998, 63(2): 95-100.
[43] ZHANG Hui,GONG Maoguo,ZHANG Puzhao,et al. Feature-level Change Detection Using Deep Representation and Feature Change Analysis for Multispectral Imagery[J]. IEEE Geoscience and Remote Sensing Letters, 2016, 13(11): 1666-1670.
[44] ZHUANG Huifu,DENG Kazhong,FAN Hongdong,et al. Strategies Combining Spectral Angle Mapper and Change Vector Analysis to Unsupervised Change Detection in Multispectral Images[J]. IEEE Geoscience and Remote Sensing Letters, 2016, 13(5): 681-685.
[45] 魏立飛, 鐘燕飛, 張良培, 等. 遙感影像融合的自適應(yīng)變化檢測(cè)[J]. 遙感學(xué)報(bào), 2010, 14(6): 1196-1211.
WEI Lifei, ZHONG Yanfei, ZHANG Liangpei, et al.Adaptive Change Method of Remote Sensing Image Fusion[J]. Journal of Remote Sensing, 2010, 14(6): 1196-1211.
[46] GONG Maoguo,ZHOU Zhiqiang,MA Jingjing. Change Detection in Synthetic Aperture Radar Images Based on Image Fusion and Fuzzy Clustering[J]. IEEE Transactions on Image Processing, 2012, 21(4): 2141-2151.
[47] ZHANG Jixian, ZHANG Yonghong. Remote Sensing Research Issues of the National Land Use Change Program of China[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2007, 62(6): 461-472.
[48] DENG JS,WANG K,DENG YH,et al. PCA-based Land-use Change Detection and Analysis Using Multitemporal and Multisensor Satellite Data[J]. International Journal of Remote Sensing, 2008, 29(16): 4823-4838.
[49] FALCO N,MARPU PR,BENEDIKTSSON JA. A Toolbox for Unsupervised Change Detection Analysis[J]. International Journal of Remote Sensing, 2016, 37(7): 1505-1526.
[50] ZHONG J,WANG R. Multi-temporal Remote Sensing Change Detection Based on Independent Component Analysis[J]. International Journal of Remote Sensing, 2006, 27(10): 2055-2061.
[51] NIELSEN A A,CONRADSEN K, Simpson J J. Multivariate Alteration Detection (MAD) and MAF Postprocessing in Multispectral, Bitemporal Image Data: New Approaches to Change Detection Studies[J]. Remote Sensing of Environment, 1998, 64(1): 1-19.
[52] NIELSEN A A. The Regularized Iteratively Reweighted MAD Method for Change Detection in Multi-and Hyperspectral Data[J]. IEEE Transactions on Image Processing, 2007, 16(2): 463-478.
[53] MARPU P R,GAMBA P, CANTY M J. Improving Change Detection Results of IR-MAD by Eliminating Strong Changes[J]. IEEE Geoscience and Remote Sensing Letters, 2011, 8(4): 799-803.
[54] WU Chen, DU Bo, ZHANG Liangpei. Slow Feature Analysis for Change Detection in Multispectral Imagery[J]. IEEE Transactions on Geoscience and Remote Sensing, 2014, 52(5): 2858-2874.
[55] WU Chen, ZHANG Liangpei, DU Bo. Kernel Slow Feature Analysis for Scene Change Detection[J]. IEEE Transactions on Geoscience and Remote Sensing, 2017, 55(4): 2367-2384.
[56] MAS J F. Monitoring Land-Cover Changes: A Comparison of Change Detection Techniques[J]. International Journal of Remote Sensing, 1999, 20(1): 139-152.
[57] DEROSE R J, LONG J N,RAMSEY R D. Combining Dendrochronological Data and the Disturbance Index to Assess Engelmann Spruce Mortality Caused by a Spruce Beetle Outbreak in Southern Utah, USA[J]. Remote Sensing of Environment, 2011, 115(9): 2342-2349.
[58] ELMORE A J,MUSTARD J F,MANNING S J,et al. Quantifying Vegetation Change in Semiarid Environments: Precision and Accuracy of Spectral Mixture Analysis and the Normalized Difference Vegetation Index[J]. Remote Sensing of Environment, 2000, 73(1): 87-102.
[59] LYON J G,YUAN Ding, LUNETTA R S,et al. A Change Detection Experiment Using Vegetation Indices[J]. Photogrammetric Engineering and Remote Sensing, 1998, 64(2): 143-150.
[60] ZENG Yuan, SCHAEPMAN M E,WU Bingfang, et al. Scaling-based Forest Structural Change Detection Using an Inverted Geometric-optical Model in the Three Gorges Region of China[J]. Remote Sensing of Environment, 2008, 112(12): 4261-4271.
[61] YUAN Fei, SAWAYA K E,LOEFFELHOLZ B C,et al. Land Cover Classification and Change Analysis of the Twin Cities (Minnesota) Metropolitan Area by Multitemporal Landsat Remote Sensing[J]. Remote Sensing of Environment, 2005, 98(2-3): 317-328.
[62] HUANG Zhi,JIA Xiuping,GE Linlin. Sampling Approaches for One-pass Land-use/Land-cover Change Mapping[J]. International Journal of Remote Sensing, 2010, 31(6): 1543-1554.
[63] GIL-YEPES J L,RUIZ L A,RECIO J A,et al. Description and Validation of a New Set of Object-based Temporal Geostatistical Features for Land-use/Land-cover Change Detection[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2016(121): 77-91.
[64] HUSSAIN M, CHEN Dongmei, CHENG A, et al. Change Detection from Remotely Sensed Images: From Pixel-based to Object-based Approaches[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2013(80): 91-106.
[65] BRUZZONE L, SERPICO S B. An Iterative Technique for the Detection of Land-cover Transitions in Multitemporal Remote-sensing Images[J]. IEEE Transactions on Geoscience and Remote Sensing, 1997, 35(4): 858-867.
[66] DEMIR B,BOVOLO F,BRUZZONE L. Detection of Land-cover Transitions in Multitemporal Remote Sensing Images with Active-learning-based Compound Classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2012, 50(5): 1930-1941.
[67] BAHIRAT K,BOVOLO F,BRUZZONE L,et al. A Novel Domain Adaptation Bayesian Classifier for Updating Land-cover Maps with Class Differences in Source and Target Domains[J]. IEEE Transactions on Geoscience and Remote Sensing, 2012, 50(7): 2810-2826.
[68] YU Wenjuan, ZHOU Weiqi, QIAN Yuguo, et al.A New Approach for Land Cover Classification and Change Analysis: Integrating Backdating and an Object-based Method[J]. Remote Sensing of Environment, 2016(177): 37-47.
[69] WU Chen, DU Bo, CUI Xiaohui, et al.A Post-classification Change Detection Method Based on Iterative Slow Feature Analysis and Bayesian Soft Fusion[J]. Remote Sensing of Environment, 2017(199): 241-255.
[70] LE HéGARAT-MASCLE S,OTTLé C,GUéRIN C. Land Cover Change Detection at Coarse Spatial Scales Based on Iterative Estimation and Previous State Information[J]. Remote Sensing of Environment, 2005, 95(4): 464-479.
[71] CHEN Jun, LU Miao, CHEN Xuehong, et al. A Spectral Gradient Difference Based Approach for Land Cover Change Detection[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2013(85): 1-12.
[72] GUEGUEN L, SOILLE P, PESARESI M. Change Detection Based on Information Measure[J]. IEEE Transactions on Geoscience and Remote Sensing, 2011, 49(11): 4503-4515.
[73] CELIK T. Change Detection in Satellite Images Using a Genetic Algorithm Approach[J]. IEEE Geoscience and Remote Sensing Letters, 2010, 7(2): 386-390.
[74] CELIK T. Unsupervised Change Detection in Satellite Images Using Principal Component Analysis and k-means Clustering[J]. IEEE Geoscience and Remote Sensing Letters, 2009, 6(4): 772-776.
[75] CELIK T, MA Kaikuang. Multitemporal Image Change Detection Using Undecimated Discrete Wavelet Transform and Active Contours[J]. IEEE Transactions on Geoscience and Remote Sensing, 2011, 49(2): 706-716.
[76] EISMANN M T,MEOLA J, HARDIE R C. Hyperspectral Change Detection in the Presence of Diurnal and Seasonal Variations[J]. IEEE Transactions on Geoscience and Remote Sensing, 2008, 46(1): 237-249.
[77] THEILER J. Quantitative Comparison of Quadratic Covariance-based Anomalous Change Detectors[J]. Applied Optics, 2008, 47(28): F12-F26.
[78] MEOLA J, EISMANN M T,MOSES R L,et al. Detecting Changes in Hyperspectral Imagery Using a Model-based Approach[J]. IEEE Transactions on Geoscience and Remote Sensing, 2011, 49(7): 2647-2661.
[79] THEILER J, SCOVEL C, WOHLBERG B, et al. Elliptically Contoured Distributions for Anomalous Change Detection in Hyperspectral Imagery[J]. IEEE Geoscience and Remote Sensing Letters, 2010, 7(2): 271-275.
[80] WU Chen, ZHANG Liangpei, DU Bo. Hyperspectral Anomaly Change Detection with Slow Feature Analysis[J]. Neurocomputing, 2015(151): 175-187.
[81] ACITO N,MATTEOLI S,ROSSI A,et al. Hyperspectral Airborne “Viareggio 2013 Trial” Data Collection for Detection Algorithm Assessment[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2016, 9(6): 2365-2376.
[82] LIU Sicong, BRUZZONE L,BOVOLO F,et al. Hierarchical Unsupervised Change Detection in Multitemporal Hyperspectral Images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2015, 53(1): 244-260.
[83] LIU Sicong, BRUZZONE L,BOVOLO F,et al. Sequential Spectral Change Vector Analysis for Iteratively Discovering and Detecting Multiple Changes in Hyperspectral Images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2015, 53(8): 4363-4378.
[84] 武辰, 杜博, 張良培. 基于獨(dú)立成分分析的高光譜變化檢測(cè)[J]. 遙感學(xué)報(bào), 2012, 16(3): 545-561.
WU Chen, DU Bo, ZHANG Liangpei. Hyperspectral Change Detection Based on Independent Component Analysis[J]. Journal of Remote Sensing, 2012,16(3): 545-561.
[85] ERTURK A,PLAZA A. Informative Change Detection by Unmixing for Hyperspectral Images[J]. IEEE Geoscience and Remote Sensing Letters, 2015, 12(6): 1252-1256.
[86] ERTüRK A,IORDACHE M D,PLAZA A. Sparse Unmixing with Dictionary Pruning for Hyperspectral Change Detection[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2017, 10(1): 321-330.
[87] ERTüRK A,IORDACHE M D,PLAZA A. Sparse Unmixing-based Change Detection for Multitemporal Hyperspectral Images[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2016, 9(2): 708-719.
[88] LIU Sicong,BRUZZONE L,BOVOLO F,et al. Unsupervised Multitemporal Spectral Unmixing for Detecting Multiple Changes in Hyperspectral Images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2016, 54(5): 2733-2748.
[89] TANG Yuqi,ZHANG Liangpei,HUANG Xin. Object-oriented Change Detection Based on the Kolmogorov-Smirnov Test Using High-resolution Multispectral Imagery[J]. International Journal of Remote Sensing, 2011, 32(20): 5719-5740.
[90] DESCLéE B, BOGAERT P, DEFOURNY P. Forest Change Detection by Statistical Object-based Method[J]. Remote Sensing of Environment, 2006, 102(1-2): 1-11.
[91] HUO Chunlei, ZHOU Zhixin, LU Hanqing, et al. Fast Object-level Change Detection for VHR Images[J]. IEEE Geoscience and Remote Sensing Letters, 2010, 7(1): 118-122.
[92] LU Ping, STUMPF A, KERLE N, et al. Object-oriented Change Detection for Landslide Rapid Mapping[J]. IEEE Geoscience and Remote Sensing Letters, 2011, 8(4): 701-705.
[93] RASI R,BEUCHLE R,BODART C,et al. Automatic Updating of an Object-based Tropical Forest Cover Classification and Change Assessment[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2013, 6(1): 66-73.
[94] WANG Biao,CHOI S,BYUN Y,et al. Object-based Change Detection of very High Resolution Satellite Imagery Using the Cross-sharpening of Multitemporal Data[J]. IEEE Geoscience and Remote Sensing Letters,2015,12(5): 1151-1155.
[95] TAN Kun,JIN Xiao,PLAZA A,et al. Automatic Change Detection in High-resolution Remote Sensing Images by Using a Multiple Classifier System and Spectral-spatial Features[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2016, 9(8): 3439-3451.
[96] TANG Yuqi, ZHANG Liangpei. Urban Change Analysis with Multi-sensor Multispectral Imagery[J]. Remote Sensing, 2017, 9(3): 252.
[97] WEN D,HUANG Xin,ZHANG Liangpei,et al. A Novel Automatic Change Detection Method for Urban High-resolution Remotely Sensed Imagery Based on Multi-index Scene Representation[J]. IEEE Transactions on Geoscience and Remote Sensing, 2016, 54(1): 609-625.
[98] HUO Chunlei,CHEN Keming,DING Kun,et al. Learning Relationship for Very High Resolution Image Change Detection[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2016, 9(8): 3384-3394.
[99] LEI Zhen,FANG Tao,HUO Hong,et al. Bi-temporal Texton Forest for Land Cover Transition Detection on Remotely Sensed Imagery[J]. IEEE Transactions on Geoscience and Remote Sensing, 2014, 52(2): 1227-1237.
[100] PACIFICI F,DEL FRATE F. Automatic Change Detection in Very High Resolution Images With Pulse-coupled Neural Networks[J]. IEEE Geoscience and Remote Sensing Letters, 2010, 7(1): 58-62.
[101] ZHONG Yanfei,LIU Wenfeng,ZHAO Ji,et al.Change Detection Based on Pulse-coupled Neural Networks and the NMI Feature for High Spatial Resolution Remote Sensing Imagery[J]. IEEE Geoscience and Remote Sensing Letters, 2015, 12(3): 537-541.
[102] Lü Pengyuan,ZHONG Yanfei,ZHAO Ji,et al.Change Detection Based on a Multifeature Probabilistic Ensemble Conditional Random Field Model for High Spatial Resolution Remote Sensing Imagery[J]. IEEE Geoscience and Remote Sensing Letters, 2016, 13(12): 1965-1969.
[103] TANG Yuqi, HUANG Xin, ZHANG Liangpei. Fault-tolerant Building Change Detection from Urban High-resolution Remote Sensing Imagery[J]. IEEE Geoscience and Remote Sensing Letters, 2013, 10(5): 1060-1064.
[104] HUANG Xin, ZHU Tingting, ZHANG Liangpei, et al.A Novel Building Change Index for Automatic Building Change Detection from High-resolution Remote Sensing Imagery[J]. Remote Sensing Letters, 2014, 5(8): 713-722.
[105] GUEGUEN L,PESARESI M,EHRLICH D,et al.Urbanization Detection by a Region Based Mixed Information Change Analysis between Built-up Indicators[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2013, 6(6): 2410-2420.
[106] HUANG Xin, ZHANG Liangpei, ZHU Tingting.Building Change Detection From Multitemporal High-resolution Remotely Sensed Images Based on a Morphological Building Index[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2014, 7(1): 105-115.
[107] SOFINA N,EHLERS M. Building Change Detection Using High Resolution Remotely Sensed Data and GIS[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2016, 9(8): 3430-3438.
[108] SINHA P, KUMAR L.Independent Two-step Thresholding of Binary Images in Inter-Annual Land Cover Change/No-Change Identification[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2013(81): 31-43.
[109] BRUZZONE L, PRIETO D F.Automatic Analysis of the Difference Image for Unsupervised Change Detection[J]. IEEE Transactions on Geoscience and Remote Sensing, 2000, 38(3): 1171-1182.
[110] MELGANI F, BAZI Y.Markovian Fusion Approach to Robust Unsupervised Change Detection in Remotely Sensed Imagery[J]. IEEE Geoscience and Remote Sensing Letters, 2006, 3(4): 457-461.
[111] BAZI Y, MELGANI F, AL-SHARARI H D. Unsupervised Change Detection in Multispectral Remotely Sensed Imagery with Level Set Methods[J]. IEEE Transactions on Geoscience and Remote Sensing, 2010, 48(8): 3178-3187.
[112] HAO Ming,SHI Wenzhong,ZHANG Hua,et al.Unsupervised Change Detection with Expectation-maximization-based Level Set[J]. IEEE Geoscience and Remote Sensing Letters, 2014, 11(1): 210-214.
[113] GONG Maoguo,ZHANG Puzhao,SU Linzhi,et al.Coupled Dictionary Learning for Change Detection from Multisource Data[J]. IEEE Transactions on Geoscience and Remote Sensing, 2016, 54(12): 7077-7091.
[114] HANSEN M C,LOVELAND T R.A Review of Large Area Monitoring of Land Cover Change Using Landsat Data[J]. Remote Sensing of Environment, 2012(122): 66-74.
[115] JIN Suming, YANG Limin, ZHU Zhe, et al.A Land Cover Change Detection and Classification Protocol for Updating Alaska NLCD 2001 to 2011[J]. Remote Sensing of Environment, 2017(195): 44-55.
[116] HUANG Xin, WEN Dawei, LI Jiayi, et al.Multi-level Monitoring of Subtle Urban Changes for the Megacities of China Using High-resolution Multi-view Satellite Imagery[J]. Remote Sensing of Environment, 2017(196): 56-75.
[117] HUANG Xin, ZHANG Liangpei, ZHU Tingting.Building Change Detection from Multitemporal High-resolution Remotely Sensed Images Based on a Morphological Building Index[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2014, 7(1): 105-115.
[118] COPPIN P R, BAUER M E. Digital Change Detection in Forest Ecosystems with Remote Sensing Imagery[J]. Remote Sensing Reviews, 1996, 13(3-4): 207-234.
[119] SHENG Yongwei, SONG Chunqiao, WANG Jida, et al.Representative Lake Water Extent Mapping at Continental Scales Using Multi-temporal Landsat-8 Imagery[J]. Remote Sensing of Environment, 2016(185): 129-141.
[120] YANG Renmin,AN Ru,WANG Huilin,et al.Monitoring Wetland Changes on the Source of the Three Rivers from 1990 to 2009, Qinghai, China[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2013, 6(4): 1817-1824.
[121] NIELSEN E M,PRINCE S D,KOELN G T. Wetland Change Mapping for the U.S.Mid-Atlantic Region Using an Outlier Detection Technique[J]. Remote Sensing of Environment, 2008, 112(11): 4061-4074.
[122] MICHALEK JL,WAGNER TW,LUCZKOVICH JJ,et al.Multispectral Change Vector Analysis for Monitoring Coastal Marine Environments[J]. Photogrammetric Engineering and Remote Sensing, 1993, 59(3): 381-384.
[123] SONNENSCHEIN R,KUEMMERLE T,UDELHOVEN T, et al.Differences in Landsat-based Trend Analyses in Drylands due to the Choice of Vegetation Estimate[J]. Remote Sensing of Environment, 2011, 115(6): 1408-1420.
[124] LATIFOVIC R,POULIOT D. Monitoring Cumulative Long-term Vegetation Changes over the Athabasca Oil Sands Region[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2014, 7(8): 3380-3392.
[125] BAGAN H, TAKEUCHI W, KINOSHITA T, et al.Land Cover Classification and Change Analysis in the Horqin Sandy Land from 1975 to 2007[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2010, 3(2): 168-177.
[126] GROSS J E,GOETZ S J,CIHLAR J. Application of Remote Sensing to Parks and Protected Area Monitoring: Introduction to the Special Issue[J]. Remote Sensing of Environment, 2009, 113(7): 1343-1345.
[127] BINDSCHADLER R A,SCAMBOS T A,CHOI H, et al.Ice Sheet Change Detection by Satellite Image Differencing[J]. Remote Sensing of Environment, 2010, 114(7): 1353-1362.
[128] KOLTUNOV A,Ustin SL.Early Fire Detection Using Non-linear Multitemporal Prediction of Thermal Imagery[J]. Remote Sensing of Environment, 2007, 110(1): 18-28.
[129] BRUNNER D,LEMOINE G,BRUZZONE L. Earthquake Damage Assessment of Buildings Using VHR Optical and SAR Imagery[J]. IEEE Transactions on Geoscience and Remote Sensing, 2010, 48(5): 2403-2420.
[130] BOVOLO F, BRUZZONE L.A Split-based Approach to Unsupervised Change Detection in Large-size Multitemporal Images: Application to Tsunami-damage Assessment[J]. IEEE Transactions on Geoscience and Remote Sensing, 2007, 45(6): 1658-1670.
[131] LI Lin, USTIN S L,LAY M. Application of AVIRIS Data in Detection of Oil-induced Vegetation Stress and Cover Change at Jornada, New Mexico[J]. Remote Sensing of Environment, 2005, 94(1): 1-16.
[132] ABUELGASIM A A,ROSS W D,GOPAL S,et al.Change Detection Using Adaptive Fuzzy Neural Networks: Environmental Damage Assessment after the Gulf War[J]. Remote Sensing of Environment, 1999, 70(2): 208-223.
[133] MA Guorui, SUI Haigang, LI Pingxiang, et al.A Kernel Change Detection Algorithm in Remote Sense Imagery[C]∥Proceedings of 2006 IEEE International Conference on Geoscience and Remote Sensing Symposium. Denver, CO, USA: IEEE, 2006:220-224.
[134] ZHAO Bei, ZHONG Yanfei, ZHANG Liangpei.A Spectral-structural Bag-of-features Scene Classifier for Very High Spatial Resolution Remote Sensing Imagery[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2016(116): 73-85.
[135] ZHANG Fan,DU Bo,ZHANG Liangpei.Scene Classification via a Gradient Boosting Random Convolutional Network Framework[J]. IEEE Transactions on Geoscience and Remote Sensing, 2016, 54(3): 1793-1802.
[136] WU Chen, ZHANG Lefei, ZHANG Liangpei.A Scene Change Detection Framework for Multi-temporal Very High Resolution Remote Sensing Images[J]. Signal Processing, 2016(124): 184-197.
[137] CARDILLE J A,FORTIN J A. Bayesian Updating of Land-cover Estimates in a Data-rich Environment[J]. Remote Sensing of Environment, 2016(186): 234-249.
[138] LING Feng, LI Wenbo, DU Yun, et al.Land Cover Change Mapping at the Subpixel Scale with Different Spatial-resolution Remotely Sensed Imagery[J]. IEEE Geoscience and Remote Sensing Letters, 2011, 8(1): 182-186.
[139] WANG Qunming,ATKINSON P M,SHI Wenzhong.Fast Subpixel Mapping Algorithms for Subpixel Resolution Change Detection[J]. IEEE Transactions on Geoscience and Remote Sensing, 2015, 53(4): 1692-1706.
[140] ZHANG Puzhao, GONG Maoguo, SU Linzhi, et al.Change Detection Based on Deep Feature Representation and Mapping Transformation for Multi-spatial-resolution Remote Sensing Images[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2016(116): 24-41.
[141] LIU Jia,GONG Maoguo,QIN Kai,et al.A Deep Convolutional Coupling Network for Change Detection Based on Heterogeneous Optical and Radar Images[J]. IEEE Transactions on Neural Networks and Learning Systems, 2017.DOI: 10.1109/TNNLS.2016.2636227.(in Press)
(責(zé)任編輯:張艷玲)
Advance and Future Development of Change Detection for Multi-temporal Remote Sensing Imagery
ZHANG Liangpei1,WU Chen2
1. State Key Laboratory of Information Engineering in Surveying,Mapping,and Remote Sensing,Wuhan University,Wuhan 430079,China; 2. International School of Software,Wuhan University,Wuhan 430079,China
Change detection in multi-temporal remote sensing images can be widely applied in monitoring ecosystem changes,and tracking urban developments,thus is extremely useful to study the interaction between human and natural environment.With the development of new remote sensing technology,change detection has attracted more and more interests in dealing with multi-temporal hyperspectral and high-resolution remote sensing images.In this review,the recent advances are introduced in four aspects:pre-processing,change detection method,thresholding and accuracy assessment.Then,the main applications for change detection are summarized.And finally,the future developments of change detection are discussed.
change detection;remote sensing imagery;multi-temporal data;high-resolution image;hyperspectral image
The National Natural Science Foundation of China(Nos. 61601333;41431175);The Natural Science Foundation of Hubei Province of China(No. 2016CFB245);The Fundamental Research Funds for the Central Universities(No. 2042016kf0034)
ZHANG Liangpei(1962—),male,PhD,professor,majors in the processing,analysis,and application of remote sensing imagery.
張良培,武辰.多時(shí)相遙感影像變化檢測(cè)的現(xiàn)狀與展望[J].測(cè)繪學(xué)報(bào),2017,46(10):1447-1459.
10.11947/j.AGCS.2017.20170340.
ZHANG Liangpei,WU Chen.Advance and Future Development of Change Detection for Multi-temporal Remote Sensing Imagery[J]. Acta Geodaetica et Cartographica Sinica,2017,46(10):1447-1459. DOI:10.11947/j.AGCS.2017.20170340.
P237
A
1001-1595(2017)10-1447-13
國(guó)家自然科學(xué)基金(61601333;41431175);湖北省自然科學(xué)基金(2016CFB245);中央高?;究蒲袠I(yè)務(wù)費(fèi)專項(xiàng)基金(2042016kf0034)
2017-06-21
修回日期: 2017-08-03
張良培(1962—),男,博士,教授,研究方向?yàn)檫b感影像處理、分析與應(yīng)用。
E-mail: zlp62@whu.edu.cn