**3. Overview of existing methods for water body mapping**

In the majority of algorithms for water body mapping a spectral band in the NIR spectral region plays an important role due to the high absorption of water and resulting high

On the Use of Airborne Imaging Spectroscopy Data for the

features are present

time.

where *green* is a green band and *NIR* is a NIR band

where *green* is a green band and *MIR* is a middle infrared band

Automatic Detection and Delineation of Surface Water Bodies 7

Fig. 3. True colour composite of an AISA image of Helgoland, Germany, with (b) histogram of the NDWI, (c) Water mask by threshold 0 (red line in histogram) on the NDWI; (d) Water mask by threshold 0.13 (green line in histogram) on the NDWI. In image c the water body (bottom left side) is almost totally included in the water mask but many urban features are so, too. In image d some parts of the water body are already lost but still some urban

*green NIR NDWI*

*green MIR MNDWI*

In addition to the spectral-based approaches object-oriented methods have been developed for water body mapping (e.g. Xiao & Tien, 2010). However, since these methods use size and shape features they have to be adjusted individually for each application and can not be used for mapping ponds, rivers and coastal waters with the same configuration at the same

 

*green NIR* 

> 

*green MIR* 

(1)

(2)

contrast in NIR bands to many other surface types. However, Manavalan et al. (1993) found that optimal cut-of gray values for individual spectral bands have to be carefully adjusted and are varying between different images. Band ratios or spectral indices are often used to mitigate spectral differences between images and also to enhance the contrast between surface types. Consequently, indices like the NDWI (McFeeters, 1996) (Equation 1) and MNDWI (Xu, 2006) (Equation 2) have been developed. Basically, the authors suggest a default threshold value of zero for these indices, i.e. gray values greater than zero represent water pixels. However, the comparative study of Ji et al. (2009) showed that an image and landscape specific adjustment of threshold values can improve results. Therefore, these methods are not fully suitable for automation. Further, NDWI shows high false positives in build-up areas (Xu, 2006). Xu developed the MNDWI to enhance the separation between water and built-up areas using Landsat ETM+ images. However, in high spatial resolution images there is no single spectral profile for the class "built-up areas" (Roessner *et al.*, 2011) and many man-made materials have positive NDWI and/or MNDWI values (Fig. 2 and Tab. 1). This is also true for shadow over non-vegetated areas. Fig. 3 shows that indices like the NDWI are not suitable for water body mapping in urban areas using high spatial resolution images since no threshold value can be found for which both, false positives and false negatives are low.

Fig. 2. Reflectance spectra of man-made materials with positive NDWI and/or MNDWI values. The gray bars indicate Landsat TM bands which are typically taken for calculating the NDWI and MNDWI. The spectra were collected from the test site Potsdam


Table 1. Corresponding index values of the spectra in Fig. 2

contrast in NIR bands to many other surface types. However, Manavalan et al. (1993) found that optimal cut-of gray values for individual spectral bands have to be carefully adjusted and are varying between different images. Band ratios or spectral indices are often used to mitigate spectral differences between images and also to enhance the contrast between surface types. Consequently, indices like the NDWI (McFeeters, 1996) (Equation 1) and MNDWI (Xu, 2006) (Equation 2) have been developed. Basically, the authors suggest a default threshold value of zero for these indices, i.e. gray values greater than zero represent water pixels. However, the comparative study of Ji et al. (2009) showed that an image and landscape specific adjustment of threshold values can improve results. Therefore, these methods are not fully suitable for automation. Further, NDWI shows high false positives in build-up areas (Xu, 2006). Xu developed the MNDWI to enhance the separation between water and built-up areas using Landsat ETM+ images. However, in high spatial resolution images there is no single spectral profile for the class "built-up areas" (Roessner *et al.*, 2011) and many man-made materials have positive NDWI and/or MNDWI values (Fig. 2 and Tab. 1). This is also true for shadow over non-vegetated areas. Fig. 3 shows that indices like the NDWI are not suitable for water body mapping in urban areas using high spatial resolution images since no threshold value can be found for which both, false positives and

MIR

Spectral profiles of selected surface types

500 1000 1500 2000

Fig. 2. Reflectance spectra of man-made materials with positive NDWI and/or MNDWI values. The gray bars indicate Landsat TM bands which are typically taken for calculating

> Copper 0.28 0.10 Plastic -0.13 0.01 Shadow 0.03 -0.10 PVC 0.03 0.20 Zinc 0.09 -0.17

the NDWI and MNDWI. The spectra were collected from the test site Potsdam

Table 1. Corresponding index values of the spectra in Fig. 2

**Surface type NDWI MNDWI** 

Wavelength [nm]

false negatives are low.

green

1000

2000

3000

Reflectance

 [%\*100]

NIR

Fig. 3. True colour composite of an AISA image of Helgoland, Germany, with (b) histogram of the NDWI, (c) Water mask by threshold 0 (red line in histogram) on the NDWI; (d) Water mask by threshold 0.13 (green line in histogram) on the NDWI. In image c the water body (bottom left side) is almost totally included in the water mask but many urban features are so, too. In image d some parts of the water body are already lost but still some urban features are present

$$\text{NNDVI} = \frac{\left(green - \text{NIR}\right)}{\left(green + \text{NIR}\right)} \tag{1}$$

where *green* is a green band and *NIR* is a NIR band

$$\text{MNDVDVI} = \frac{\left(green - MIR\right)}{\left(green + MIR\right)}\tag{2}$$

where *green* is a green band and *MIR* is a middle infrared band

In addition to the spectral-based approaches object-oriented methods have been developed for water body mapping (e.g. Xiao & Tien, 2010). However, since these methods use size and shape features they have to be adjusted individually for each application and can not be used for mapping ponds, rivers and coastal waters with the same configuration at the same time.

On the Use of Airborne Imaging Spectroscopy Data for the

consecutively remove false positives (sections 4.2 and 4.3).

Histogram of NIR spectral mean image (Helgoland)

Histogram of NIR spectral mean image (Berlin)

Reflectance [%]

Reflectance [%]

Fig. 5. Histograms (left: full, right: subset) of the NIR spectral mean images of two test sites

spectral region from 440 nm to 2500 nm.

**4.1 Masking potential water pixels** 

Number of pixels

Number of pixels

(top: Helgoland, bottom: Berlin)

Automatic Detection and Delineation of Surface Water Bodies 9

488 to 60 spectral bands, respectively. The mean spectral sampling interval of the analyzed datasets is 2.3 nm for "Döberitzer Heide" and 4.6 nm for "Helgoland". The HyMap sensor is an airborne VNIR-SWIR whiskbroom scanner with 16 bit radiometric resolution consisting of four detector modules with mean spectral sampling intervals of 15 nm (VIS and NIR), 13 nm (SWIR1) and 17 nm (SWIR2) (Cocks *et al.*, 1998). The 128 spectral bands cover the

Water detection is a trivial task as long as there are no other dark surfaces present in the image. Unfortunately, the most prominent spectral characteristic of water pixels – water pixels are very dark – also applies to a couple of other surfaces such as dark rocks (e.g., lava, basalt) or bituminous roofing materials and especially to pixels covered by shadow. To account for this, we developed a two-step approach that firstly masks low albedo pixels as potential water pixels (section 4.1) and secondly applies a process of elimination to

Masking of potential water pixels is done by thresholding a spectral mean image of all NIR bands between 860 nm and 900 nm of a sensor. As pointed out before water absorbs most of the incident energy in the NIR spectral region exhibiting a high brightness contrast to the majority of other surfaces. However, since every scene is different a scene-specific threshold has to be found. This is done automatically based on the histogram of the NIR spectral mean image (Fig. 5). After finding the histogram peak of low albedo surfaces (first local

> Number of pixels

Number of pixels Subset for polynomial approximation (Helgoland)

Subset for polynomial approximation (Berlin)

Reflectance [%]

Reflectance [%]
