Logo GISdevelopment.net

GISdevelopment > Proceedings > ACRS > 2002


1989 | 1990 | 1991 | 1992 | 1994 | 1995 | 1996 | 1997 | 1998 | 1999 | 2000 | 2002
Sessions

GIS, GPS & Data Integration

Land Use Land Cover

Hazard Mitigation and Disaster Management

Photogrammetry

Forestry

Earth Observation from Space

Mountain Environment and Mapping

Data processing, Algorithm and Modelling

Urban Mapping

Hyperspectral Data Acquisition and Systems

AIT: Digital Asia

SAR / InSAR

Very High Resolution Mapping

Soil and Agriculture

Water Resources

Geology / Geomorphology

Education

Ecology, Environment & Carbon Cycle

Infrastructure Planning and Management

Oceanography and Coastal Zone Monitoring

Poster Sessions

Poster 1

Poster 2

Poster 3



ACRS 2002


Ecology, Environment & Carbon Cycle


A Study on Land Use and Environment in Tonza City area in Hainan Province, China: Twenty Years of Change



2.2 General Methodology
The following is the brief description of methods u sed in this study.

GPS field work
GPS fieldwork was carried out in Septemer 2001, for collecting ground control points for geometric correction and for collecting training data for the classification of satellite data. Differential GPS was carried out using Trimble GPS system using Geo Explorer III and 4600LS as a base, and using the software Pathfinder for differential correction.

Geometric correction
Since the topographic map that was available did not have coordinate information, the geometric correction of the satellite data was carried out using the geometric coordinates of the GCP’s calculated from our GPS field survey. The topographic map was also registered and rectified using these GCP’s so that topographic map and the satellite data are matched in the same map projection.

Classification of satellite data
Training data were collected during field survey, which were later used to make supervised maximum likelihood classification of the satellite data. This was relatively simple for Landsat ETM data, since there was little change in land use condition at the time of the ground truthing (September 2001) compared to the time the data was acquired (January 1999). For the areas where some change was noticed, information from local people was used to ascertain the land cover during the time of satellite data acquisition. However, classification of MSS data was more complex since this data was from 1980. For this, ground truth data were used for known areas, based on the knowledge acquired from the local people who knew of the condition of land cover at that time. Knowledge from our extensive field work at various times was employed for this purpose. Using such ground truth information, a supervised maximum likelhihood classification was performed both on Landsat ETM as well as MSS data to derive the land use/cover classes in 1980 and 1999. The satellite data have been classified into six simplified categories of mature (dense) forest, shrub (sparse forest), grass, paddy, water and bare soil (including built up area). The MSS data for 1980 contains come cloud (about 14% of total area). Therefore, for comparison of land use between 1980 and 1999, only the area, which is not covered by cloud in both images, is considered.

Digital Elevation Modelling
A digital elevation model was prepared based on the topographic map. Even though this topographic map was not very detailed, it could still be used for studying the general topographic conditions of the area and linking the land use/cover with respect to the terrain conditions. For this purpose, maps for height ranges and slope were computed from the DEM, which was then used for relating the land use with topography.

Page 3 of 6
| Previous | Next |

Applications | Technology | Policy | History | News | Tenders | Events | Interviews | Career | Companies | Country Pages | Books | Publications | Education | Glossary | Tutorials | Downloads | Site Map | Subscribe | GIS@development Magazine | Updates | Guest Book