地球信息科学学报
地毬信息科學學報
지구신식과학학보
GEO-INFORMATION SCIENCE
2013年
2期
280-288
,共9页
范俊甫%马廷%周成虎%周玉科
範俊甫%馬廷%週成虎%週玉科
범준보%마정%주성호%주옥과
DMSP-OLS%经验阈值修正%环渤海城市群%建成区%空间格局
DMSP-OLS%經驗閾值脩正%環渤海城市群%建成區%空間格跼
DMSP-OLS%경험역치수정%배발해성시군%건성구%공간격국
DMSP-OLS%thresholds correction%Bohai Rim%urban built-up area%spatial pattern
从 DMSP-OLS 数据提取城市区域的经验阈值法存在固有的缺陷,经验阈值对不同的空间区域不具备通用性,不适用的经验阈值,将导致城区面积提取具有较大误差,可采用统计数据对经验阈值进行修正以降低误差。本文在大时空尺度条件下以少量样本城市的统计数据对经验阈值方法进行了修正;另采用Elvidge二次多项式模型对DMSP-OLS时间序列数据进行了校正。在此基础上选取总斑块数量、景观总面积、平均斑块大小、最大斑块面积比、斑块密度、景观形状指数、总边界长度、平均边界密度和斑块平均回旋半径共9个景观生态学指标,采用FRAGSTATS 3.3软件计算分析了1992-2010年环渤海区域城市发展的空间格局变化特征。结果表明:1992-2010年间,环渤海区域城市化进程持续较快发展,城市建成区总面积增加了2.14倍,平均城市建成区面积增加了76%,提取到的城市斑块平均回旋半径增加了约26.5%,并且城市景观斑块的复杂度明显上升,可检测到的城镇总个数增加了82%。但是,每100km2内孤立的城市景观斑块数却降低了约76%;大城市持续扩张的相对速度慢于中小城市,城市区域边界密度和整体城市分布景观格局破碎度逐渐降低,核心城市与周围卫星城市的景观斑块的联通性逐步增加。
從 DMSP-OLS 數據提取城市區域的經驗閾值法存在固有的缺陷,經驗閾值對不同的空間區域不具備通用性,不適用的經驗閾值,將導緻城區麵積提取具有較大誤差,可採用統計數據對經驗閾值進行脩正以降低誤差。本文在大時空呎度條件下以少量樣本城市的統計數據對經驗閾值方法進行瞭脩正;另採用Elvidge二次多項式模型對DMSP-OLS時間序列數據進行瞭校正。在此基礎上選取總斑塊數量、景觀總麵積、平均斑塊大小、最大斑塊麵積比、斑塊密度、景觀形狀指數、總邊界長度、平均邊界密度和斑塊平均迴鏇半徑共9箇景觀生態學指標,採用FRAGSTATS 3.3軟件計算分析瞭1992-2010年環渤海區域城市髮展的空間格跼變化特徵。結果錶明:1992-2010年間,環渤海區域城市化進程持續較快髮展,城市建成區總麵積增加瞭2.14倍,平均城市建成區麵積增加瞭76%,提取到的城市斑塊平均迴鏇半徑增加瞭約26.5%,併且城市景觀斑塊的複雜度明顯上升,可檢測到的城鎮總箇數增加瞭82%。但是,每100km2內孤立的城市景觀斑塊數卻降低瞭約76%;大城市持續擴張的相對速度慢于中小城市,城市區域邊界密度和整體城市分佈景觀格跼破碎度逐漸降低,覈心城市與週圍衛星城市的景觀斑塊的聯通性逐步增加。
종 DMSP-OLS 수거제취성시구역적경험역치법존재고유적결함,경험역치대불동적공간구역불구비통용성,불괄용적경험역치,장도치성구면적제취구유교대오차,가채용통계수거대경험역치진행수정이강저오차。본문재대시공척도조건하이소량양본성시적통계수거대경험역치방법진행료수정;령채용Elvidge이차다항식모형대DMSP-OLS시간서렬수거진행료교정。재차기출상선취총반괴수량、경관총면적、평균반괴대소、최대반괴면적비、반괴밀도、경관형상지수、총변계장도、평균변계밀도화반괴평균회선반경공9개경관생태학지표,채용FRAGSTATS 3.3연건계산분석료1992-2010년배발해구역성시발전적공간격국변화특정。결과표명:1992-2010년간,배발해구역성시화진정지속교쾌발전,성시건성구총면적증가료2.14배,평균성시건성구면적증가료76%,제취도적성시반괴평균회선반경증가료약26.5%,병차성시경관반괴적복잡도명현상승,가검측도적성진총개수증가료82%。단시,매100km2내고립적성시경관반괴수각강저료약76%;대성시지속확장적상대속도만우중소성시,성시구역변계밀도화정체성시분포경관격국파쇄도축점강저,핵심성시여주위위성성시적경관반괴적련통성축보증가。
The photoelectric amplification characteristics of Operational Linescan System (OLS) sensors on board of the Defense Meteorological Satellite Program’s (DMSP) satellites make the instruments sensitive to low visible lights in the night which can distinguish the differences of light signals between urban and rural ar-eas. Remotely sensed nighttime lights datasets derived from the DMSP-OLS sensors have been extensively ap-plied to assess and monitor the process of urbanization and human activities, which has become an important da-ta source for studies on regional urbanization and human activities. Methods used to extract urban built-up areas from DMSP-OLS data, such as empirical global thresholding-based methods and the sudden detection method, cannot avoid their own defects. The experience thresholding values are not universal in different regions and the sudden detection method cannot be applied in large scales. In this study, we corrected the experience threshold-ing values by introducing statistical data of some sample cities in the research area which combined with a cali-bration process to DMSP-OLS time serial data for extracting urban built-up area from satellite-based nighttime light data at large temporal and spatial scales. Nine landscape metrics: the number of patches (NP), the landscape total area (TA), the mean patch size (MPS), the largest patch index (LPI), the patches density of per hundred km2 (PDh), the landscape shape index(LSI), the total edge length(TE), the edge density(ED) and the radius of gyra-tion (GYRATE) are calculated by the FRAGSTATS3.3 software to analysis the spatial pattern change characteris-tics of urban area in Bohai Rim. The study showed that from 1992 to 2010, the urbanization in Bohai Rim experi-enced a continuing and rapid process. In this region, the total urban built-up areas expanded for 2.14 times, the average built-up area of cities increased for 76%, the gyrate of extracted urban patches expanded about 26.5%which suggested that the complexity of urban patch shapes were increased. The amount of detected urban patch-es got 82% increase but the number of isolate cities in each 100 km2 were decreased by about 76% which im-plied that the expansion of traditional cities was the dominant factor of the area increasing rather than continuous-ly emerging towns. The expansions of metropolises were slower than small cities, and the overall landscape frag-mentation degree was decreased gradually with the trend of urban area connection between core cities and their exurbs.