A Systematic Survey on Different Topographic Correction Techniques for Rugged Terrain Satellite Imagery

IJECT Vol. 4, IssuE spl - 5, July - sEpT 2013 ISSN : 2230-7109 (Online) | ISSN : 2230-9543 (Print)

w w w . i j e c t . o r g 14 InternatIonal Journal of electronIcs & communIcatIon technology

A Systematic Survey on Different Topographic Correction Techniques for Rugged Terrain Satellite Imagery

1Sartajvir Singh, 2Dr. Rajneesh Talwar1Dept. of ECE, Chitkara University, Himachal Pradesh, India

2Chandigarh Engineering College, Punjab, India

AbstractIn remote sensed data, topographic corrections are very important to overcome the topographic effects such as shadow effects. These topographic effects occurs especially in case of rugged terrain because sun facing slope represents maximum reflectance values and opposite side slope represents minimum reflectance values. These topographic effects leads to an inaccurate classification into different land cover classes due to different topographic angles. Therefore, effective removal or minimization of topographic effects is necessary in satellite imagery data of rugged terrain to extract the maximum information. Different topographic correction techniques have been developed to reduce topographical effects but each topographic technique has a different impact in different situation. In paper report, all recently developed different topographic correction techniques for land use/cover have been discussed with their feature and limitation for rugged terrain. Further research with comparative analysis of each method is needed on a global basis to derive guidelines on which method performs best under which situation.

KeywordsTopographic Effects, Satellite Image Preprocessing, Topographic Correction Techniques, Slope Match

I. IntroductionRemote sensed data is only a practical mean to provide the earth imagery for environment monitoring, forecasting of natural hazards etc. But many satellite imageries are affected by atmospheric distortions such as geometric distortion and topographic distortions. In order to provide proper interpretation and analysis of the satellite imagery, these distortions must be overcome or minimized. This paper report generally focus on topographic effects that occurs due to variations in observed radiance from an inclined surface as compared to spectro-radiometric responses from a horizontal surface as a function of effective incident angle and emergent angle [1-2].A number of topographic correction techniques have been developed to reduce the topographic effects in satellite imagery of mountainous region. It is also found [3-4] that some well-developed topographic correction algorithms are problematic in operation, and some of them cannot perform well enough in some situation. Every topographic correction technique has different impact on different situation. In 1971, topographic correction method involves simplest band ratio method in which single band is divided by another band [5-6]. Another category of topographic correction in which the Hyper-Spherical Direction Cosine Transformation (HSDC) projects measurement vectors onto a hyper-sphere.In this paper report, all topographic correction based on Digital Elevation Model (DEM), have been discussed. The topographic correction techniques are generally divided in three categories. First category involves lambertian topographic correction approaches in which surface reflects irradiation in all directions equally and direct irradiance is considered such as cosine [7-9], cosine-T [8],

C-correction [8], cosine-C [7], smooth C [8], SCS (sun canopy senor), SCS plus [4] etc. Second category involves non-lambertian topographic correction approaches which diffuse irradiance is modeled by means of constants wavelength dependent assessment of the constants for each band separately reflection characteristics depending on land cover individual constants such as Minneart [10] etc. and third category involves empirical approaches in which an empirically determined calibration coefficient (C) such as two stage normalization [11], Slope match [12] etc. The main aim of this paper report to the review of different topographic correction techniques for rugged terrain. This article is divided into five sections, following this introduction of topographic correction categories, an overview preprocessing methodology is presented in the second section, followed by topographic correction with its types in subsections. In the fourth section, discussion of prior studies on topographic correction techniques and general conclusions are drawn in Section fifth.

II. Preprocessing MethodologyThe Preprocessing of remote sensed imageries is very essential for various correction such as geometric corrections, radiometric correction and in addition this, topographic correction is also very important for satellite image interpretation & further analysis. These topographic corrections plays a very significant role especially in case of rugged terrain satellite imagery. The essential preprocessing methodology for satellite imagery interpretation has shown in fig. 1.

III. Topographic CorrectionThe topographic corrections are generally refer to compensation of topographic effects. The topographic correction must be applied on reflectance image before perform on further analysis. All topographic correction techniques discussed in paper are based on Digital Elevation Model (DEM). The three categories of topographic correction have been discussed in following section.

Raw Satellite Imagery

Geometric Corrections

Digital Elevation Model

Generation

Illumination Angle

Estimation of Reflectance

Imagery

Estimation of coefficients

Topographic Corrected Reflectance Imagery

Radiometric Corrections

Aspect Imagery

Slope Imagery

Topographic Corrections

Further Analysis e.g. Change Detection,

monitoring

Fig. 1: Methodology of Satellite Image Preprocessing

IJECT Vol. 4, IssuE spl - 5, July - sEpT 2013

w w w . i j e c t . o r g InternatIonal Journal of electronIcs & communIcatIon technology 15

ISSN : 2230-7109 (Online) | ISSN : 2230-9543 (Print)

A. Lambertian Correction ApproachesThe lambertian correction approaches are based on assumption the surface reflects incident solar energy uniformly in all directions equally. All Lambertain based topographic correction techniques are discussed in following subparts.

1. Cosine CorrectionThe cosine topographic correction technique fall into lambertian category in which surface is assumed to be a perfect diffuse reflector and having the same amount of reflectance in all view directions. Under the assumption of lambertian surfaces, the cosine correction [3, 13] has been extensively used to correct for illumination variations [3, 14]. It is calculated according to following equation.

= (1) In above equation, spectral reflectance for horizontal surface spectral reflectance observed over the inclined terrain is denoted by and respectively. The solar zenith angle is denoted by

The Illumination (IL) is denoted by [3, 7, 15].

2. Cosine-C CorrectionIn cosine-C correction technique, the average IL conditions has been considered. It is a modified version cosine law lambertian technique. This technique is wavelength independent, since the correction is based on the same factor for all the bands as reported in [16]. It is calculated according to following equation.

= (2)

In above equation, represents the mean of illumination of study area and rest of parameters are same as equation (1)

3. Cosine-b CorrectionIn cosine-b correction technique, a coefficient is added to cosine law which is given by equation (3). Based on an examination of satellite imagery, a linear relationship exists between spectral reflectance observed over the inclined terrainand illumination in the form (4), called regression equation. In equation (4), and

” parameters represents slope and intercept.

= (3)

(4)

4. Smooth C correctionSmooth C [3] is a variation of C-correction method based on a smoothed IL value. Most methods produce an overcorrection in those pixels where IL is low. Therefore, a variation in the calculation of the IL was carried out, by smoothing the original slope with a smoothing factor of 3, 5, and 7 as proposed by [3]. The slope of surface and slope factor is denoted by in following equation and , respectively.

= arc (5)

5. SCS (Sun-Canopy-Sensor)The SCS technique [17] involves normalizing the sunlit area within a pixel without changing the sun and sensor positions or the orientation, geometry, and structure of the canopy. Sun Canopy Sensor (SCS) geometry should more effectively remove the topographic effect on snow cover area. It is calculated according

to following equation.

= (6)

6. SCS+CIt is a modified sun-canopy-sensor [4] and introduced because the SCS model overestimates radiance, a problem common with earlier photometric models that do not characterize diffuse radiation properly under some terrain orientation. Soenen et al. [4] proposed the SCS+C correction where the moderator C is derived using (4) and (8) but within the improved physical context of the SCS model. This addition is intended to be an improvement to the SCS correction in a similar way as the C-correction improves on the cosine correction.

= (7)

(8)

7. C-Huang WeiC-Huang Wei method is used for topographic correction under Lambertain methods [18] proposed by Huang Wei. In following equation, Minimum value of spectral reflectance and IL is denoted by and , respecticvely.

= ( ) + (9)

B. Non-Lambertian Correction ApproachesIt is reported in [19] lambertian assumptions are very unrealistic on rugged region which leads to either underestimation or overestimation of physical parameters significantly both on sunlit slopes as well as the slopes away from the Sun. This problem is overcome by considering non-lambertian assumption. The non-lambertian approaches diffuse irradiance is modeled by means of constants wavelength dependent assessment of the constants for each band separately reflection characteristics depending on land cover individual constants.

1. Correction-CIn correction-c technique, is introduced to the cosine correction model as an additive term in (1) and rests of the parameters same.

= (10)

In above equation, correction coefficient ” is calculated according to following equation (8), and ” can be calculated using equation (4).

2. Minneart CorrectionThe Minneart topographic correction technique [20-21] is non-Lambertian and empirically derives for each band a Minneart constant using (8). Minneart constants are estimated by calculating the slope of the regression line as in (4) and (8). The value of Minneart constant can be calculated according to equation (11).

= (11)



Moreover, the Minneart method did not consistently normalize radiance for the same incidence angle value from band to band as reported in [19].

3. Pixel-Based Minneart CorrectionAs reported in [22], Minneart correction technique is most frequently used for topographic correction, but a single global Minneart value cannot effectively reduce topographic effects on the remotely sensed data, especially in the areas with steep slopes. This paper explores the method to develop a pixel-based Minneart coefficient image based on the established relationship between Minneart coefficients and topographic slopes. The evaluation of topographic correction results is conducted based on texture images, which are derived with the homogeneity texture measure. The homogeneity is calculated with (12) and (13)

(12)

(13)In above equation, is the value in the cell (row and column ) of the moving window, and n is the number of rows or columns. In order to explore the effects of slope on topographic correction results, a mean homogeneity value for each slop is calculated with (14).

(14)In above equation, is the mean value of homogeneity at the slope , and is the number of samples in the slope . Higher values indicate more homogenous.

C. Empirical Correction ApproachesThe Empirical correction approaches involves an empirically determined calibration coefficient The Empirical topographic correction techniques have been discussed in following sub section.

1. Two Stage NormalizationBefore implementing the two stage normalization topographic correction technique, the reflectance imagery must be normalized using the illumination model. Then the single stage normalization is calculated according to following equation.

= (15)

In above equation, normalized reflectance values for image pixel in wave band. It is reported in [19] the further improvement

in topographic correction by Civco [7] from the satellite imagery in second stage normalization process in which an empirically estimated calibration coefficient second stage process is computed according to following equation.

(16)In above equation, represents uncorrected reflectance imagery. The parameter north aspect and north aspect after first stage normalization is represented by and . The second stage normalization correction must be implemented on uncorrected reflectance image according to following equation to generate topographic corrected imagery:

= × (17)

2. Slope MatchA recently developed slope match topographic correction technique [12] involves certain modifications to Civco’s method [7] and performed the correction in two stages because they observed that Civco’s [7] not provide the well results in shadow areas as per reported in [9]. The final reflectance for topographic correction is estimated using (12) proposed by Nichol and others [12].

= × (18)

In above equation, represents the topographically corrected reflectance imagery. The spectral reflectance on the tilted surface is represented by . The parameters and represents maximum and minimum reflectance, respectively and estimated from topographically uncorrected reflectance image, is mean value of illumination on the south aspect. The parameter

is normalization coefficient for different satellite bands and estimated using equation given in the literature [12].

(19)In above equation, representes mean reflectance on sunlit slopes after first stage normalization, represents mean reflectance on shady slopes in uncorrected reflectance imagery and is the mean reflectance value on shady slope after first stage normalization.

3. VECAA simple empirical based topographic correction technique, the variable Empirical coefficient Algorithm (VECA) , was developed [18] using theoretical and statistics analysis of the radiance values of remotely sensed data acquired for rugged terrain and the cosine of the solar illumination angle . In case of topographic effect being involved, the statistical relationship between imagery radiance and topographic variable can be calculated according to following equation.

(20)In above equation, represents the mean of the corrected pixel radiance value of the imagery and represents the mean of the uncorrected pixel radiance value of the imagery. A parameter is an adjustment factor. The rugged terrain located on the North Slope represents an adjustment factor as , and on the South slope when surface is the horizontal . An adjustment factor decrease with the increase of means are inversely proportional to each other. So an adjustment parameter controls the topographic corrections, that why it is called Variable Empirical Coefficient Algorithm.

IV. Results and DiscussionAll above mentions techniques are based on DEM. The simplest lambertian cosine correction tended to be over-corrected, with slopes facing away from the sun appearing brighter than sun-facing slopes due to diffuse sunlight being relatively more influential on the shady slope. An improvement in cosine law is done by introducing a coefficient in cosine law, called cosine-c which is a wavelength independent. Further improvement in Cosine law is done by introducing a coefficient in cosine

IJECT Vol. 4, IssuE spl - 5, July - sEpT 2013

w w w . i j e c t . o r g InternatIonal Journal of electronIcs & communIcatIon technology 17

ISSN : 2230-7109 (Online) | ISSN : 2230-9543 (Print)

law, called cosine-b which is based on regression equation. But these methods are generally overcorrect the results. The SCS technique is more appropriate than other corrections but in some terrain orientations, SCS faces an overcorrection problem similar to other simple photometric functions. To overcome this problem, SCS+C correction has been introduced which accounts for diffuse atmospheric irradiance based on the C-correction.This hypothesis of lambertian reflectance characteristics considers only direct irradiance although the proportion of diffuse skylight can be relatively high, particularly in weakly illuminated areas. Several non-lambertian techniques have been developed to account for the shortcomings of the lambertian assumption. The non-lambertian Minneart technique showed, for all bands and images, the highest residual (or introduced) topographic correlation of corrected radiance data. Moreover, the Minneart method did not consistently normalize radiance for the same incidence angle value from band to band. Minneart method has indicated the drawback of using a single global k value in topographic correction, but no effective approaches have been developed. This research develops a new approach for calculation of a pixel-based k image based on the relationship between k and slope and uses a homogeneity approach to evaluate the topographic correction result.Afterwards, empirical based topographic correction have been developed which involves two-stage normalization, slope match, VECA etc. The two-stage normalization technique it does not equalize the reflectance imagery between the south and north aspect instead of this consider only mean IL. So slope match technique corrected reflectance imagery is normalized to the mean illumination level of pixels on the north aspect rather than the overall mean illumination value of the entire image. The slope matching topographic correction technique adjusts the brightness between northern and southern slopes and provides true quantitative retrieval of spectral reflectance, especially in shadow area.

IV. ConclusionIn this article, different topographic correction techniques for rugged terrain have been reviewed and summarized with their merits and demerits. It is found that each topographic perform different in different situation, some methods over estimate and under estimate the parameters but still some methods perform superb under different situation so further research is needed with imagery on a global basis to derive guidelines on which method performs best under which situation.

References[1] Holben B.N., Justice C.O.,“The topographic effect on spectral

response from nadir-pointing sensors”, Photogrammetric Engineering and Remote Sensing, Vol. 46, No. 9, 1980, pp. 1191–1200.

[2] Justice C.O., Wharton S.W., Holben B.N.,“Application of digital terrain data to quantify and reduce the topographic effect on Landsat data”, International Journal of Remote Sensing, Vol. 2, No. 3, 1981, pp. 213–230.

[3] Riano, E. Chuvieco, J. Salas, I. Aguado,“Assessment of different topographic corrections in Landsat-TM data for mapping vegetation types”, IEEE Trans. Geosci. Remote Sensing, Vol. 41, No. 5, 2003, pp. 1056–1061.

[4] Scott A. Soenen, Derek R. Peddle,“SCS+C: A Modified Sun-Canopy-Sensor Topographic Correction in Forested Terrain”, IEEE Transactions on Geoscience and Remote Sensing, Vol. 43, No. 9, 2005, pp. 2148-2159.

[5] Crane .B.,“Preprocessing techniques to reduce atmospheric and sensor variability multispectral scanner data”, In Proceedings of the Seventh International Symposium on Remote Sensing of Environment (Ann Arbor: Environmental Research Institute of Michigan), pp. 1345-1355, 1971.

[6] Vincent R.K.,“Spectral ratio imaging methods for geological remote sensing from aircraft and utilization of remotely sensed data”, In Proceedings of the American Society of Photogrammetry Management and Utilization of Remote Sensing Data Conference A, Anson (Ed.), Sioux Falls, South Dakota, pp. 377-397 (Virginia: American Society of Photogrammetry), 1973.

[7] L. Civco,“Topographic normalization of Landsat Thematic Mapper digital imagery”, Photogramm. Eng. Remote Sens., Vol. 55, 1989, pp. 1303–1309.

[8] P. M. Teillet, B. Guindon, D. G. Goodenough,“On the slope-aspect correction of multispectral scanner data”, Can. J. Remote Sens., Vol. 8, No. 2, 1982, pp. 84–106.

[9] J. Dozier, D. Marks,“Snow mapping and classification from Landsat Thematic Mapper data", Ann. Glaciol., Vol. 9, 1987, pp. 97–103.

[10] M. Minnaert,"The reciprocity principle in lunar photometry”, Astrophys. J., Vol. 93, 1941, pp. 403–410.

[11] Colby J. D.,“Topographic normalization in rugged terrain”, Photogramm. Eng. Remote Sens. Vol. 57, 1991, pp. 531–537.

[12] J. Nichol, L. K. Hang, W. M. Sing,“Empirical correction of low sun angle images in steeply sloping terrain: A slope matching technique”, Int. J. Remote Sens. Vol. 27(3–4), 2006, pp. 629–635.

[13] B. Holben, C. Justice,“The topographic effect on spectral response from nadir pointing sources”, Photogramm. Eng. Remote Sens., Vol. 46, No. 9, 1980, pp. 1191–1200.

[14] P. Meyer, K. I. Itten, T. Kellenberger, S. Sandmeier, R. Sandmeier,“Radiometric corrections of topographically induced effects on Landsat TM data in an alpine environment”, ISPRS J. Photogramm. Remote Sens., Vol. 48, No. 4, 1993, pp. 17–28.

[15] Y. Kawata, S. Ueno, A. Ohtani,“The surface albedo retrieval of mountainous forest area from satellite MSS data”, Appl. Math. Comput., Vol. 69, 1995, pp. 41–59.

[16] P. M. Teillet, B. Guindon, D. G. Goodenough,“On the slope-aspect correction of multispectral scanner data”, Can. J. Remote Sens., Vol. 8, No. 2, 1982, pp. 84–106.

[17] D.Gu, A. Gillespie,“Topographic normalization of Landsat TM images of forest based on subpixel sun-canopy-sensor geometry”, Remote Sens. Environ., Vol. 64, 1998, pp. 166–175.

[18] Yongnian Gao, Wanchang Zhang,“A simple empirical topographic correction method for ETM+ imagery”, International Journal of Remote Sensing, Vol. 30, No. 9, 2009, pp. 2259 – 2275.

[19] V. D. Mishra, J. K. Sharma, K. K. Singh, N. K. Thakur, M. Kumar,“Assessment of different topographic corrections in AWiFS satellite imagery of Himalaya terrain”, J. Earth Syst. Sci., Vol. 118, No. 1, 2009, pp. 11–26.

[20] J. A. Smith, T. L.Tzeu Lie Lin, K. J. Ranson,“The Lambertian assumption and Landsat data”, Photogramm. Eng. Remote Sens., Vol. 46, 1980, pp. 1183–1189.

[21] Massimo Vincini, Ermes Frazzi,“Multitemporal Evaluation of Topographic Normalization Methods on Deciduous Forest TM Data”, IEEE Transactions on Geoscience and Remote



Sensing, Vol. 41, No. 11, 2003, pp. 2586 – 2590.[22] Dengsheng Lu, Hongli Ge, Shizhen He, Aijun Xu, Guomo

Zhou, Huaqiang Du,“Pixel-based Minnaert Correction Method for Reducing Topographic Effects on a Landsat 7 ETM_ Image”, Photogrammetric Engineering & Remote Sensing Vol. 74, No. 11, 2008, pp. 1343–1350.

Sartajvir Singh, is an Assistant Professor at the Department of Electronics and Communication, Chitkara University, Himachal Pradesh, India. His research interest is in the area of Digital Image Processing.

Dr. Rajneesh Talwar, is presently working as Principal at Chandigarh Engineering College, Landran Punjab, India. He has professional experience of more than ten years in teaching and research.

Documents

A Systematic Survey on Different Topographic Correction Techniques for Rugged Terrain Satellite Imagery