УДК Trinh Le Hung, Mai Dinh Sinh, Nguyen Van Bien LAND SURFACE TEMPERATURE RETRIEVAL FROM LANDSAT ULTISPECTRAL IMAGE

Transcription

1 УДК Trinh Le Hung, Mai Dinh Sinh, Nguyen Van Bien LAND SURFACE TEMPERATURE RETRIEVAL FROM LANDSAT ULTISPECTRAL IMAGE Статья посвящена решению актуальной проблемы определения поверхностной температуры городской территории Ханоя по данным теплового инфракрасного изображения для обнаружения аномальных участков перегрева, который может принимать меры по решению эффекта городского теплового острова Ключевые слова: Вьетнам, многозональная съёмка, тепловая инфракрасная съемка, поверхностная температура. Introduction. The thermal environment in urban areas is characterized by the heat island phenomenon affected energy demand, human health and environmental conditions. The urban air temperature is gradually rising in all cities in the world. One of the causes is the drastic reduction in the greenery area in cities. The distinguished climatic condition termed urban heat island is developing in the rapidly urbanized cities. Ground-based observations reflect only thermal condition of local area around the station. We can not establish the number of meteorological stations with expected density. Nowadays thermal remote sensing has been used over urban areas to assess the urban heat island. To estimate the thermal condition of land surface by satellite, it is necessary to find the relationship between the surface temperature and land cover type. The LANDSAT data with 60m (ETM+), 120m (TM) and 100m (LANDSAT 8) spatial resolution of thermal infrared band enables users to define the more detailed surface temperature. LANDSAT images are widely used to observe and model the biophysical characteristics of the land surface. In addition to the development of Land use/cover maps band 6 of the LANDSAT imagery is useful for deriving the surface temperature. Materials and methodology. The study area is located at the North Vietnam (Hanoi city and Vinh Phuc province). Hanoi is second largest cities in Vietnam with more than 4 million populations. The high economic growth and abundant employment opportunities caused influx of labor immigration. The urban heat islands effect is clearly noticeable in the inner city districts of Hanoi city with up to 10 degrees higher than the average temperature of the surrounding areas Fig. 1. LANDSAT ETM+ images of the study area for the year 2007 and 2009 Trinh Le Hung, Mai Dinh Sinh, Nguyen Van Bien,

2 ISSN Технические науки To calculating land surface temperature we used LANDSAT ETM+ satellite image on 08 November 2007 and 05 November 2009 (fig. 1). The Enhanced Thematic Mapper (ETM+) on board LANDSAT 7 is a multispectral radiometric sensor that records eight bands of data with varying spectral and spatial resolutions (30m spatial resolution for red, green, blue, near infrared and two bands of medium infrared; 60m for thermal infrared; and a 15m panchromatic band). Conversion of the digital number to spectral radiance. Image processing started with geometric and radiometric correction. Radiometric correction done by converted the digital number value in LANDSAT thermal band (band 6, band 6.1 and band 6.2) to radiance value. Method to convert digital number to radiance value is shown in equation 1: Lmax Lmin L ( DN DN min) Lmin DN max DN min (1) where, L λ - spectral radiance watts/(m*m * ster * µm); DN digital number; Lmin spectral radiance which is correlate with DNmin watts/(m*m * ster * µm); Lmax spectral radiance which is correlate with DNmax watts/(m*m * ster * µm); DNmin = 1, minimum value of DN; DNmax = 255, maximum value of DN. Value of Lmax, Lmin for image LANDSAT ETM, ETM+ Table 1. Band Satellite/Sensor Lmax Lmin 6.1 LANDSAT7 /ETM+ High gain LANDSAT7 /ETM +Low gain LANDSAT ETM, ETM Conversion of the spectral radiance to brightness temperature Method to convert radiance value to effective temperature value is shown in equation 2: where, K T 2 B K ln(1 1 ) L T - brightness temperature; B L λ - spectral radiance watts/(m*m * ster * µm); K1 = W/(m*m*ster*µm), calibration const; K2 = W/(m*m*ster*µm), calibration const. (2) Estimation of land surface emissivity. An alternative, operative procedure is to obtain the land surface emissivity image from the NDVI. The method proposed obtains the emissivity values from the NDVI considering different cases: a) NDVI < 0.2. In this case, the pixel is considered as bare soil and then a constant value for the emissivity is assumed, typically of b) NDVI > 0.5. Pixels with NDVI values higher than 0.5 are considered as fully vegetated and then a constant value for the emissivity is assumed, typically of c) 0.2 < NDVI < 0.5. In this case, the pixels is composed by a mixture of bare soil and vegetation, and the emissivity is calculated according to the following equation: 59

3 (3) v P v s( 1 P v ) 0.04P v 0.95 where εv is the vegetation of the emissivity and εs is the soil emissivity. Pv is the vegetation proportion obtained according to the following equation: 2 NDVI NDVI P min (4) v NDVImax NDVI min NDVI is normalized difference vegetation index, which is calculated according to the following equation: NIR RED NDVI (5) NIR RED Calculation land surface temperature. Method to calculate land surface temperature is show in equation 6: LST T. 1 ( TB B )*ln (6) where T - brightness temperature (K B 0 ), λ wavelength (11.5 µm), ε land surface emissivity,, h Plank s constant (6, J.sec), c velocity of light (2, m/sec), σ Stefan Boltzmann s constant, which is equal to 5, Wm -2 K -4. Results and discussions. Basing on the NDVI values of different land use classes, emissivity image was prepared by assigning emissivity values as 0.95 for building areas and 0.99 for vegetation areas. The emissivity image is shown in Fig.2 below Fig. 2. Emissivity image of the study area for the year 2007 (left) and 2009 (right) 60

4 ISSN Технические науки From brightness temperature and land surface emissivity images, the final land surface temperature image was obtained by developing a program in C++. The final land surface temperature image is shown in the Fig.3 below. From the land surface temperature image , it was observed that highest temperature of about 29,1 0 C exist at urban building areas and other impervious areas and lowest temperature of about 9,9 0 C are existing at vegetative areas. With land surface temperature image , the highest temperatures was C and the lowest temperatures about C Fig. 3 Land Surface Temperature image of the study area for the year2007 and 2009 Conclusitons. Hanoi is the second largest city in Vietnam and it is experiencing a rapid urbanization. With urbanization most of the land surface is covered with concrete, asphalt and other such impervious materials. This leads a variety of urban environmental issues like increase in runoff increase in land surface temperatures. The cities are experiencing more heat than the surrounding rural areas mainly due lack of vegetative cover. In this study an attempted is made to identify relationship between land use/cover and land surface temperature. Remote sensing has the capability of monitoring such changes, extracting the change in information from satellite data. From the land surface temperature images it is clearly understood that surface temperature is more in urban area compared to rural areas. Also the correlation study shows that the land surface temperature is strongly and negatively correlated with NDVI. This information can be used in monitoring the dynamics of land use resulting out of changing demands of increasing population and associated issues like urban heat island. References 1. K. Sundara Kumar, P. Udaya Bhaskar, K. Padmakumari (2012), Estimation of land surface temperature to study urban heat island effect using LANDSAT ETM+ image, International journal of Engineering Science and technology, Vol. 4, No. 2, pp O.R. Garcia Cueto, E. Jauregui Ostos, D. Toudert, A. Tejeda Martinez (2007), Detection of the urban heat island in Mexicali and its relationship with land use, Atmosfera 20(2), pp LANDSAT 7 Science data users Handbook, National aeronautics and space administration (NASA), 186 pp. 61

5 4. Van de Griend A.A., Owen M. (1993). On the relationship between thermal emissivity and the normalized difference vegetation index for natural surface // International journal of remote sensing 14, pp Valor E., Caselles V. (1996). Mapping land surface emissivity from NDVI. Application to European African and South American areas // Remote sensing of Environment, 57, pp Статья поступила в редакцию ЧИНЬ ЛЕ ХУНГ кандидат технических наук, Технический институт им. Ле Куи Дон, Ханой, Вьетнам. МАЙ ДИНЬ ШИНЬ магистр, Технический институт им. Ле Куи Дон, Ханой, Вьетнам. НГУЕН ВАН БЬЕН магистрант, Технический институт им. Ле Куи Дон, Ханой, Вьетнам. 62