| Impervious surface is an important indicator in the study of urbansprawl and urban environmentusing remote sensing technology. It plays also an important role in many applications, such as urban land cover monitoring, analysis of urban sprawl, spatial distributionassessment of population, estimation of non-point source pollution, urban planning and environmental assessment. It is important in sustainable development of an urban system too. As a typical city in the hilly-gullies Loess region of northwest China and the capital of Gansu Province, Lanzhou city is a key industrial base, trade center, scientific research center, and high-level education center in northwest China. In thegreat course of economic reform of China, the economy in this area has made great strides forward in the last decade, the cityacceleratedly sprawls in alldirection, and impervious surface changescorrespondingly. This article analyzes the classification method of impervious surface, the rules of and driving forces of impervious surface changes at the city based on Landsat Thematic Mapper (TM) data and object-based image analysis (OBIA). and demonstrates the potential of geospatial information technology with respect of scientific and socio-economic values.In this paper, the author explored the classification methods of impervious surface in urban based on time series of Landsat images in20years withobject-base image analysis methodology. Based on the analysis of the area, distribution, dynamic changes and change rules of impervious surface, the author analyzes the reasons for impervious surface changes considering natural factors and human activities.The main conclusions are as follows:(1) For extraction of impervious surface information, existing Normalized Difference Impervious Surface Index (NDISI) is concerned essentially with the cities in South China, there are a large number of ecological land use, such as farmland, vegetation and water. But around the city is mostly mountainous and hilly, bare soil is found widely in the hilly and gully region of the Loess Plateau in northwest China. There are great difficulties in information extraction because of the poor separability of the spectral signature between impervious surface and bare soil, so NDISI index do not apply to this study area. This paper, based on the Normalized Difference Impervious Surface Index (NDISI)index, create a Modified Normalized Difference Impervious Surface Index (MNDISI) which is more consistent with real circumstance of the study area. The result shows that impervious surface information enhancement is more significant, the results are satisfactory using this index.(2) Object-based image analysis method can extract impervious surface and the basic set of other city land cover types effectively in the study area. The rich spectral and spatial characteristics are good basic for extraction of impervious surface information. The results of classification and accuracy assessment show that the object-based image analysis method can avoid "salt-and-pepper" phenomenon which commonly happens in classical pixel-based image analysis.(3) This article evaluated the image segmentation results using the discrepancy measures. We used three metric indices. Potential Segmentation Error (PSE), Number-of-Segments Ratio (NSR), and Euclidian Distance2, to select the optimal segmentation parameter combination. The three metric indices focus not only geometric relationships but also arithmetic relationships between reference polygons and corresponding segments.(4) After analyzing various characteristics of impervious surface and the basic set of other city land cover types in study area, this article extract and classify the information according to six main features that consist of Modified Normalized Difference Impervious Surface Index (MNDISI), the Modified Normalized Difference Water Index (MNDWI), Soil Adjusted Vegetation Index (SAVI). and the average ratio of reflectivity of different land cover types in the near infrared and short-wave infrared band in the nearest neighbor classifier of object-based image analysis method. The accuracy evaluation of impervious surface information classification results in the multi-images show that producer’s accuracy, user accuracy, overall accuracy and Kappa coefficient were90%,88%,89%and0.86. Show higher classification accuracy of this method, and we can extract the impervious surface information more effectively.(5) According to research, the area of impervious surface has been growing rapidly, which from98.46hm2in1993to129.13hm2in2000. to162.03hm2in2011. The percentage of all the basic set of city ground cover types from17.16%, to22.50%, to28.24%. growth range is bigger. At the same time, the areas of water and vegetation ecological land were reduced in different degrees. Within the last few decades, due to the government has been concerned about the ecological construction, the area of ecological land decrease rate decreased. Moreover, in the past twenty years the coverage of bare soil area has little change in the study area. (6) The area of increased impervious surface throughout the study area, and it change fast. Increased impervious surface areas are mainly concentrated in coastal areas of the Yellow River Xigu District, Cuijiadatan that is binding region of Xigu District and Qilihe District, highway201North in Arming District, Hewan in Anning District, Huangjiatan in Anning District, Houwuquan village in Qilihe District, Jiuzhou Development Zone, Yantan village, Donggangzhen Village in Chengguan District, east of the region along State Road312, and so on from1993to2000. From2000to2011increased impervious surface areas are mainly concentrated in coastal areas of Beitan and Datan that are binding region of Xigu District and Qilihe District, the university district in Anning District, resident area like Anning Courtyard, the region along State Road212and Lanzhou-Linxia highway in Qilihe District, Houwuquan and Lujia village in Qilihe District, Jiuzhou development Zone, the region of Yantan village eastern, the region along Liugouhe-Zhonghexiang highway and Lianyungang-Huoerguosi highway, and so on.(7) By comparing the brightness temperature map and the corresponding extraction map of impervious surface information in different times, we know that the change trend of high temperature region and extension region of impervious surface is a one-to-one relationship, that they show synchronism growth. The industrial estate, the commercial area and living area with dense population had an important influence in the formation of high-temperature zone. Coverage rate of ecological land such as vegetation and water is inversely proportional to intensity of heat island, which can weaken the urban heat-island effect.(8) The driving forces which caused the change of impervious surface at Lanzhou include natural factors and human factors. Natural factors mainly include the basic conditions such as the geographical position and topographical features, which restrict the layout of city spatial structure and the selection of functional area of Lanzhou. Human factors include economic development, population increase, policy and planning and the development of road traffic. However, in the study area human factors play a more prominent role in affecting impervious surface. |
PDF Full Text Request