**Land Use/Cover Classification Techniques Using Optical Remotely Sensed Data in Landscape Planning**

Onur Şatr and Süha Berberoğlu

*Cukurova University, Agriculture Faculty, Department of Landscape Architecture, Turkey* 

## **1. Introduction**

20 Landscape Planning

Weeks, P. and Mehta, S. (2004). Managing People and Landscapes: IUCN's Protected Area

Yücel M. (1995). Protected Areas and Planning. Publication No. 104, C¸ ukurova University

Categories, J. Hum. Ecol, (16) 4: 253-263.

Press, Adana, Turkey, 255 pp. (in Turkish).

The observed biophysical cover of the earth's surface, termed land-cover is composed of patterns that occur due to a variety of natural and human-derived processes. On the other hand Land-use is human activity on the land, influenced by economic, cultural, political, historical, and land-tenure factors. Remotely-sensed data (i.e., satellite or aerial imagery) can often be used to define land-use through observations of the land-cover (Brown, et al., 2000; Karl & Maurer, 2010). Up-to-date land-use information is of critical importance to planners, scientists, resource managers, and decision makers.

Optical remote sensing (RS) plays a vital role about defining LUC (land use/cover) and monitoring interactions between nature and human activities. Additionally, RS provides time, energy and cost saving. Today, optical RS data such as satellite sensor images and aerial photos are used widely to detect LUC dynamics. LUC mapping outcomes are used for global, regional, local mapping, change detection, landscape planning and driving landscape metrics.

RS image classification is a complex process and requires consideration of many factors. The major steps of image classification may include i) determination of a suitable classification system, ii) image preprocessing iii) selection of training samples, iv) selection of suitable classification approaches and post-classification processing, and v) accuracy assessment. Additionally, the user's need, scale of the study area, economic condition, and analyst's skills are important factors influencing the selection of remotely sensed data, the design of the classification procedure, and the quality of the classification results (Lu and Weng 2007).

LUC mapping has been used for various purposes in landscape planning and assessment such as, deriving landscape metrics (Southworth et al., 2010, Huang et al., 2007), landscape monitoring (Özyavuz et al., 2011, Berberoglu and Akin 2009), LUC change modeling (e.g., SLUETH (Clarke, 2008)), agricultural studies (agricultural policy environmental extender model (APEX) (Gassman et al., 2010); soil water assessment tool (SWAT) (Betrie et al. 2011)) and environmental processes (revised universal soil loss equation (RUSLE) (Renard et al., 1997)).

Land Use/Cover Classification Techniques

Using Optical Remotely Sensed Data in Landscape Planning 23

important data to define environmental strategies. CORINE system aims at collecting comparable and consistent land cover data across Europe. This information system, offers the essential elements for the applications of nature conservation, urban planning and resource management. The European nomenclature distinguishes 44 different types of land cover. Individual countries can supplement these categories with a more detailed level if they desire so. CORINE Land Cover is bridging the gap between the local (micro) and the EU (macro) scales. CORINE Land Cover is a platform of communication not only for environmental

information, but also for the policies that shape the environment (figure 1).

Fig. 1. The legend of CORINE LUC classes codes and colors (EEA 2012b).

The EUNIS habitat classification is a common reporting language on habitat types at European level, sponsored by the EEA. It originated from a combination of several habitat classifications - marine, terrestrial and freshwater. The terrestrial and freshwater classification builds upon previous initiatives, notably the CORINE biotopes classification (Devillers & Devillers-Terschuren 1991), the Palaearctic habitats classification (Devillers & Devillers-Terschuren 1996), of the EU Habitats Directive 92/43/EEC, the CORINE Land

**2.3 EUNIS habitat classification scheme** 

This chapter evaluates classification methods together with optical remote sensing data, and ancillary data integration to improve classification accuracy of LUC mapping.

#### **2. LUC classification schemes**

Standardization is one of the most discussed issues in LUC classification studies, and scientist and map developers were aware that using a common classification schemes might be more comparable and available. The first standardization works started in USA. Today there are several LUC schemes on the world according to region and scale. This chapter will discuss three largely used schemes; i) USGS (US geological survey) Anderson, ii) CORINE (Coordination of information on the environment) and iii) EUNIS (European Nature Information System) habitat schemes.

#### **2.1 USGS Anderson classification schemes**

This classification scheme was utilized within large number of models in the context of land physical dynamics and natural risk assessment. USGS classification scheme is based on James Anderson's system. This scheme is included nine main categories and four different levels (Anderson et al., 1976).

Level I is suitable for 1/250.000 – 1/150.000 scale imags like MODIS and Envisat MERIS. Level II is useful for higher spatial resolution satellite sensor images with a scale of 1:80,000. Level III is suitable for 1:20,000 to 1/80,000 scale images such as, Landsat 4-7 . Level IV is the most useful for images at scales larger than 1:20,000 (Ikonos, Kompsat, Rapid eye, Formosat, Geoeye, World view and aerial photos). Categories are designed to be adaptable to the local needs. Sample of Level I categories and forest land levels are showed in table 1.


Table 1. USGS classification scheme for level I of forest cover.

#### **2.2 CORINE classification scheme**

The European council found EEA (European environmental agency) in 1990 to search and discuss the environmental issues all around the Europe. LUC of Europe is one of the most

This chapter evaluates classification methods together with optical remote sensing data, and

Standardization is one of the most discussed issues in LUC classification studies, and scientist and map developers were aware that using a common classification schemes might be more comparable and available. The first standardization works started in USA. Today there are several LUC schemes on the world according to region and scale. This chapter will discuss three largely used schemes; i) USGS (US geological survey) Anderson, ii) CORINE (Coordination of information on the environment) and iii) EUNIS (European Nature

This classification scheme was utilized within large number of models in the context of land physical dynamics and natural risk assessment. USGS classification scheme is based on James Anderson's system. This scheme is included nine main categories and four different

Level I is suitable for 1/250.000 – 1/150.000 scale imags like MODIS and Envisat MERIS. Level II is useful for higher spatial resolution satellite sensor images with a scale of 1:80,000. Level III is suitable for 1:20,000 to 1/80,000 scale images such as, Landsat 4-7 . Level IV is the most useful for images at scales larger than 1:20,000 (Ikonos, Kompsat, Rapid eye, Formosat, Geoeye, World view and aerial photos). Categories are designed to be adaptable to the local

needs. Sample of Level I categories and forest land levels are showed in table 1.

41 Broadleaved Forest (generally deciduous) **42 Coniferous Forest**  43 MixedConifer-Broadleaved Forest

Table 1. USGS classification scheme for level I of forest cover.

**LEVEL I LEVEL II LEVEL III LEVEL IV** 

The European council found EEA (European environmental agency) in 1990 to search and discuss the environmental issues all around the Europe. LUC of Europe is one of the most

**421 Upland conifers** 

4211 White pine 4212 Red pine 4213 Jack pine 4214 Scotch pine 4215 White spruce 4219 Other

422 Lowland conifers

ancillary data integration to improve classification accuracy of LUC mapping.

**2. LUC classification schemes** 

Information System) habitat schemes.

levels (Anderson et al., 1976).

1 Urban or Built-up

2 Agricultural Land 3 Rangeland **4 Forest Land**  5 Water 6 Wetland 7 Barren Land 8 Tundra

9 Perennial Ice or

**2.2 CORINE classification scheme** 

Land

Snow

**2.1 USGS Anderson classification schemes** 

important data to define environmental strategies. CORINE system aims at collecting comparable and consistent land cover data across Europe. This information system, offers the essential elements for the applications of nature conservation, urban planning and resource management. The European nomenclature distinguishes 44 different types of land cover. Individual countries can supplement these categories with a more detailed level if they desire so. CORINE Land Cover is bridging the gap between the local (micro) and the EU (macro) scales. CORINE Land Cover is a platform of communication not only for environmental information, but also for the policies that shape the environment (figure 1).

Fig. 1. The legend of CORINE LUC classes codes and colors (EEA 2012b).

### **2.3 EUNIS habitat classification scheme**

The EUNIS habitat classification is a common reporting language on habitat types at European level, sponsored by the EEA. It originated from a combination of several habitat classifications - marine, terrestrial and freshwater. The terrestrial and freshwater classification builds upon previous initiatives, notably the CORINE biotopes classification (Devillers & Devillers-Terschuren 1991), the Palaearctic habitats classification (Devillers & Devillers-Terschuren 1996), of the EU Habitats Directive 92/43/EEC, the CORINE Land

Land Use/Cover Classification Techniques

**3. Remotely sensed data sources** 

Sensor Organization Spatial

Geoeye Geoeye ABD Pan: 0.41m

Ikonos Geoeye ABD Pan: 1m

Digital globe ABD

France

NASA ABD 17m

Alos (AVNIR2) JAXA Japan 10m 4 VNIR

Quickbird-2 Digital globe

Rapideye Rapideye AG

Spot 5 SpotIMAGE

ASTER Japan&ABD

Landsat 8\* NASA ABD Pan: 15m

hyperspectral ESA EU 17-34m

World view (2) and (3)\*

AVIRIS airborne hyperspectral

CHRIS proba

potential) are the other important factors (Table 3).

ABD 0.61m

Germany 5m

resolution

MS: 0.61m

MS: 4m

Pan: 0.5m MS: 2m at nadir

Pan:2.5m VNIR:10m SWIR:20m

VNIR: 15m SWIR: 30m TIR: 90m

MS: 30m

Using Optical Remotely Sensed Data in Landscape Planning 25

Data characteristics are the most important issue to select appropriate available one for a LUC mapping. Both airborne and spaceborne data have various spatial, radiometric, spectral and temporal resolutions. Large numbers of studies have focused on characteristics of remotely sensed data (Barnsley 1999, Lefsky and Cohen 2003). Additionally, scan width (cover size in one scene), data availability (accessibility) and lunch date (data archive

> Spectral resolution

4 VNIR bands (450- 920nm)

4 VNIR bands (450- 890)

4 VNIR bands (450- 880nm)

4 VNIR bands and red edge (440-850nm)

4 VNIR standart bands & 4 VNIR unique bands (400- 1040nm)

3 VNIR & 1 SWIR bands (490-1750)

224 VNIR & SWIR bands (400-2500nm)

4 VNIR bands 6 SWIR bands 5 TIR bands (520- 11650nm)

5 VNIR bands 2 SWIR bands 1 cirrus band (433-2300nm)

18-62 VNIR bands (400- 1100nm)

Radiometric resolution

Temporal resolution

12 bit 5.5 days at

11 bit ~3 days 15X15km 2008

11 bit 3.5 days 16.5X16.5km 2002

11 bit 3.5 days 11X11km 1999

11 bit 1-3days 16.4X16.4km 2009 (2)

8 bit 26 days 60X60km 2002

16 bit airborne 11X11km -

8 bit 16 days 185X185km Dec.

16 bit 16 days 15X15km 2002

16 days 60X60km 1999

bands 8 bit 46 days 70X70km 2006

VNIR:8 bit SWIR: 8 bit TIR: 12 bit

Imaging Swath

nadir 77X77km 2008

Lunch date

2014 (3)

2012

Cover nomenclature (Bossard et al. 2000), and the Nordic habitat classification (Nordic Council of Ministers 1994). The marine part of the classification was originally based on the BioMar classification (Connor et al 1997), covering the North-East Atlantic. The EUNIS habitat classification introduced agreed criteria for the identification of each habitat unit, while providing a correspondence with these earlier classification systems.

The habitat classification forms an integral part of the EUNIS, developed and managed by the European Topic Centre for Nature Protection and Biodiversity (ETC/NPB in Paris) for the EEA and the European Environmental Information Observation Network (EIONET). The EUNIS web application (http://eunis.eea.europa.eu) (EEA 2012a) provides access to publicly available data in a consolidated database.

The information includes:


The resulting system of classification is still somewhat transitional. Down to level 3 (terrestrial and freshwater) and level 4 (marine), EUNIS habitats are now based on physiognomic and physical attributes, together with some floristic criteria. There are 10 main habitat categories in this scheme. Coastal habitats and main categories as an example were presented in this chapter (table 2). Detailed information can be found in revised EUNIS habitat classification report of Davies et al. (2004).


Table 2. Main EUNIS habitat classes and sample levels of coastal habitats.

#### **3. Remotely sensed data sources**

24 Landscape Planning

Cover nomenclature (Bossard et al. 2000), and the Nordic habitat classification (Nordic Council of Ministers 1994). The marine part of the classification was originally based on the BioMar classification (Connor et al 1997), covering the North-East Atlantic. The EUNIS habitat classification introduced agreed criteria for the identification of each habitat unit,

The habitat classification forms an integral part of the EUNIS, developed and managed by the European Topic Centre for Nature Protection and Biodiversity (ETC/NPB in Paris) for the EEA and the European Environmental Information Observation Network (EIONET). The EUNIS web application (http://eunis.eea.europa.eu) (EEA 2012a) provides access to

Data on Species, Habitats and Sites compiled in the framework of NATURA2000 (EU

Data collected from frameworks such as EIONET, data sources or material published by

Information on Species, Habitats and Sites taken into account in relevant international

Specific data collected in the framework of the EEA's reporting activities, which also

The resulting system of classification is still somewhat transitional. Down to level 3 (terrestrial and freshwater) and level 4 (marine), EUNIS habitats are now based on physiognomic and physical attributes, together with some floristic criteria. There are 10 main habitat categories in this scheme. Coastal habitats and main categories as an example were presented in this chapter (table 2). Detailed information can be found in revised EUNIS

**B1 Coastal dunes and**

B1.1 Sand beach driftlines

the driftline

dunes

slacks B1.9 Machair

B1.2 Sand beaches above

B1.4 Coastal stable dune grassland (grey dunes) B1.5 Coastal dune heaths B1.6 Coastal dune scrub B1.7 Coastal dune woods B1.8 Moist and wet dune

B1.3 Shifting coastal

shores, including the

**sandy shores**  B2 Coastal shingle B3 Rock cliffs, ledges

supralittoral

and

ETC/NPB (formerly the European Topic Centre for Nature Conservation).

**LEVEL I LEVEL II LEVEL III** 

Table 2. Main EUNIS habitat classes and sample levels of coastal habitats.

while providing a correspondence with these earlier classification systems.

publicly available data in a consolidated database.

conventions or from International Red Lists.

habitat classification report of Davies et al. (2004).

E. Grasslands and lands dominated

H. Inland unvegetated and sparsely

I. Regularly or recently cultivated agricultural, horticultural and

J. Constructed, industrial and other

by forbs, mosses or lichens F. Heathland, scrub and tundra G. Woodland, forest and other

constitute a core set of data to be updated periodically.

Habitats and Birds Directives),

The information includes:

A. Marine habitats **B. Coastal habitats**  C. Inland surface waters D. Mires, bogs and fens

wooded land

vegetated habitats

domestic habitats

artificial habitats

Data characteristics are the most important issue to select appropriate available one for a LUC mapping. Both airborne and spaceborne data have various spatial, radiometric, spectral and temporal resolutions. Large numbers of studies have focused on characteristics of remotely sensed data (Barnsley 1999, Lefsky and Cohen 2003). Additionally, scan width (cover size in one scene), data availability (accessibility) and lunch date (data archive potential) are the other important factors (Table 3).


Land Use/Cover Classification Techniques

2011) (Figure 2).

Using Optical Remotely Sensed Data in Landscape Planning 27

ATCOR2), have been developed for radiometric and atmospheric normalization and correction (Chavez 1996, Heo and FitzHugh 2000, Hadjimitsis et al. 2004, Ozyavuz et al.

Fig. 2. (a) Non-corrected data, (b) Atmospherically corrected and haze removed Landsat

There are two basic approaches to the classification process: supervised and unsupervised classification. With supervised classification, one provides a statistical description of the manner in which expected land cover classes should appear in the imagery, and then a procedure (known as a classifier) is used to evaluate the likelihood that each pixel belongs to one of these classes. With unsupervised classification, a very different approach is used. Here another type of classifier is used to uncover commonly occurring and distinctive reflectance patterns in the imagery, on the assumption that these represent major land cover classes. The analyst then determines the identity of each class by a combination of experience and ground truth (i.e., visiting the study area and observing the actual cover types) (Eastman 2003). Three essential parts are vital in a LUC mapping in classification

In this chapter four different classifiers and approaches were evaluated in the example of Landsat TM sub-scenes recorded over Eastern Mediterranean coastal part. Methods and performances were assessed based on accuracy, capability and applicability. This assessment covered traditional (minimum distance, maximum likelihood, linear discriminant analyses), machine learning (decision tree, artificial neural network, support vector machine), fuzzy (linear mixture modeling, fuzzy c-means, artificial neural network, regression tree) and object based classifiers for LUC mapping. The summary of the

TM data using ATCOR2 model (Özyavuz et al. 2011)

**a b** 

stage; training, classifying and testing (accuracy assessment).

techniques and classifiers for various purposes were provided in table 4.

**4.2 Classification techniques** 

**4.2.1 Classifiers** 


Table 3. The most used optical sensor specifications in LUC mapping. (\*) planned missions.

#### **4. LUC mapping techniques**

Suitable remotely sensed data, classification systems, available classifier and number of training samples are prerequisites for a successful classification. Cingolani et al. (2004) identified three major problems when medium spatial resolution data are used for vegetation classifications: i) defining adequate hierarchical levels for mapping, ii) defining discrete land-cover units discernible by selected remote-sensing data, and iii) selecting representative training sites. In general, a classification system is designed based on the user's need, spatial resolution of selected remotely sensed data, compatibility with previous work, image-processing and classification algorithms, and time constraints. Such a system should be informative, exhaustive, and separable (Jensen 1996, Landgrebe 2003). In many cases, a hierarchical classification system is adopted to take different conditions into account (Lu and Weng 2007).

#### **4.1 Image pre-processing**

Image pre-processing includes geometric correction or image registration, atmospheric correction and radiometric calibration essentially. In addition, topographic correction and noise reduction may be applied if necessary. Optical images from current systems have already corrected geometrically (Landsat TM/ETM, MODIS) or can be corrected using freely available software or tools (e.g. BEAM for MERIS and CHRIS and MRT toolbox for MODIS).

Accurate geometric rectification or image registration of remotely sensed data is a prerequisite for a combination of different source data in a classification process. Many textbooks and articles have described this topic in detail (Jensen 1996, Toutin 2004). However, Geometric correction output should have the transformation rms. errors (RMSE) less than 1.0 pixel, indicating that the images are located with an accuracy of less than a pixel.

Atmospheric and radiometric corrections may not be necessary if a single image is used, but multitemporal or multisensor data are needed atmospheric and radiometric correction and calibration. A variety of methods, ranging from simple relative calibration such as, darkobject subtraction to calibration approaches based on complex models (e.g. MODTRAN, 6S,

Radiometric resolution

Temporal resolution

bands 16 bit 1 day 1150km 2002

TIR bands 10 bit 1 day 1446km <sup>1978</sup>

16 bit 16 days 7.7X185km 2000

14 bit 4 days 30X30km 2014-

12 bit 1 day 2330km 2000

Imaging Swath

Lunch date

2015

1998

Spectral resolution

220 VNIR & SWIR bands (400-2500nm)

244 VNIR & SWIR bands (420-2450)

36 VNIR & SWIR & TIR bands

Table 3. The most used optical sensor specifications in LUC mapping. (\*) planned missions.

Suitable remotely sensed data, classification systems, available classifier and number of training samples are prerequisites for a successful classification. Cingolani et al. (2004) identified three major problems when medium spatial resolution data are used for vegetation classifications: i) defining adequate hierarchical levels for mapping, ii) defining discrete land-cover units discernible by selected remote-sensing data, and iii) selecting representative training sites. In general, a classification system is designed based on the user's need, spatial resolution of selected remotely sensed data, compatibility with previous work, image-processing and classification algorithms, and time constraints. Such a system should be informative, exhaustive, and separable (Jensen 1996, Landgrebe 2003). In many cases, a hierarchical classification system

Image pre-processing includes geometric correction or image registration, atmospheric correction and radiometric calibration essentially. In addition, topographic correction and noise reduction may be applied if necessary. Optical images from current systems have already corrected geometrically (Landsat TM/ETM, MODIS) or can be corrected using freely available software or tools (e.g. BEAM for MERIS and CHRIS and MRT toolbox for MODIS). Accurate geometric rectification or image registration of remotely sensed data is a prerequisite for a combination of different source data in a classification process. Many textbooks and articles have described this topic in detail (Jensen 1996, Toutin 2004). However, Geometric correction output should have the transformation rms. errors (RMSE) less than 1.0

Atmospheric and radiometric corrections may not be necessary if a single image is used, but multitemporal or multisensor data are needed atmospheric and radiometric correction and calibration. A variety of methods, ranging from simple relative calibration such as, darkobject subtraction to calibration approaches based on complex models (e.g. MODTRAN, 6S,

pixel, indicating that the images are located with an accuracy of less than a pixel.

is adopted to take different conditions into account (Lu and Weng 2007).

Sensor Organization Spatial

hyperspectral NASA ABD 30m

EnMAP\* DLR Germany 30m

MODIS NASA ABD 250-1000m

**4. LUC mapping techniques** 

**4.1 Image pre-processing** 

MERIS ESA EU 300-1200m 15 VNIR

NOAA 15 NASA ABD 1090m 6 VNIR &

Hyperion

AVHRR

resolution

ATCOR2), have been developed for radiometric and atmospheric normalization and correction (Chavez 1996, Heo and FitzHugh 2000, Hadjimitsis et al. 2004, Ozyavuz et al. 2011) (Figure 2).

Fig. 2. (a) Non-corrected data, (b) Atmospherically corrected and haze removed Landsat TM data using ATCOR2 model (Özyavuz et al. 2011)
