Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Master → Chine → 2009 → A Study on Desertification Based on RS and GIS Methods in Dunhuang City

Lanzhou University (2009)

A Study on Desertification Based on RS and GIS Methods in Dunhuang City

Wang Xiao Yun

Titre : A Study on Desertification Based on RS and GIS Methods in Dunhuang City

Auteur : Wang Xiao Yun

Grade : Master’s Theses 2009

Université : Lanzhou University

Desertification is one of the socio-economic problems facing the world today the environment. China is one of the countries of the world most affected by desertification, caused great harm to the ecological environment and socio-economic. Remote sensing technology, the desertification monitoring provides a whole new means. Due to the complexity of land desertification, visual interpretation is still the primary means for the current desertification monitoring. Visual interpretation exists positioning, time-poor, long cycle, heavy workload, and other defects, therefore, it is necessary to establish a monitoring land desertification computer automatic classification method and meaningful. Dunhuang is a historical and cultural city, national excellent tourist city, and has a certain influence in the international arena. City within many cultural relics, unique natural landscape. However, in recent decades, the the Dunhuang territory land desertification increasingly heavier, a serious impact on the local ecological environment and people’s production life. In this regard, Wen given instructions, we must accelerate the governance of desertification. So, understanding and knowledge of the development and changes of land desertification in Dunhuang City, has important practical significance of local desertification control and environmental protection. Decision Tree Classification organic to combine a variety of information, showed a huge advantage in the remote sensing classification problems. The vegetation coverage is the main reference index of land desertification degree of classification, NDVI vegetation index is the most commonly used remote sensing estimation of vegetation coverage. Vegetation coverage based on the introduction of surface temperature and texture features can improve the classification accuracy of all levels of desertified land, better distinguish predominantly class. Therefore, Dunhuang study area, based on TM remote sensing data, the use of decision tree classification method, trying to establish a way to remote sensing quantitative inversion results (vegetation coverage and surface temperature) and texture analysis of the results of the introduction of the automatic classification process of desertification monitoring The new method, and to analyze the trend of land desertification in Dunhuang City, 20 years, to explore the distribution pattern and development changes, the main conclusions are : (1) Dunhuang City land desertification situation is quite serious. It is a visual interpretation of the results, the total area of ??desertified land in Dunhuang City, 2004 7551km 2, desertification degree of moderate and severe land desertification situation is quite serious ; desertified land is mainly distributed in the region south of the Shule and desertified land distribution associated with a particular land use patterns, causes different. (2) the Dunhuang City desertified land in the past 20 years has been in the trend of expansion, and the degree of desertification has also changed. 1987-2006, Dunhuang City of desertified land continue to expand, and after 10 years, the rate of expansion was significantly greater than the previous 10. At the same time, the degree of desertification dominant from 1987, the moderate transition to serious dominant in 1996 and 2006, the moderate and severe co-led process of desertification situation. (3) the causes of desertification are mainly geographical location and climatic factors, hydrological factors, vegetation factors and human factors ; Countermeasures to increase in accordance with the law and sand control efforts, the development of sandy area industries, and promote the scientific development of the sandy area, follow the laws of nature and the economy law, to take comprehensive measures to prevent desertification policy incentives to improve public awareness of prevention, strengthen coordination, supervision and inspection and reasonable solution to the utilization of water resources and strengthen desertification dynamic monitoring. (4) the use of remote sensing the quantitative inversion results (vegetation cover, surface temperature) and texture analysis results into decision tree classification, it is feasible to achieve the automatic classification of land desertification monitoring. Comparability of results of the decision tree classification method, but the accuracy is significantly lower than the man-machine interactive visual interpretation.

Mots clés : Dunhuang City Desertification Vegetation coverage Surface temperature Texture analysis Decision tree classification

Présentation (Dissertationtopic )

Page publiée le 22 mai 2013, mise à jour le 11 mars 2018