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

Accueil du site → Master → Chine → 2008 → Study on Extraction Method of Arid Zone Wetlands Information Based on Remote Sensing Technic

Université du Xinjiang (2008)

Study on Extraction Method of Arid Zone Wetlands Information Based on Remote Sensing Technic


Titre : Study on Extraction Method of Arid Zone Wetlands Information Based on Remote Sensing Technic

Auteur : 张芳;

Grade : Master’s Theses 2008

Université : Université du Xinjiang

Résumé partiel
Wetlands are natural comprehensive systems created by interactions of water and land, which are located at transitional zones between terraneous ecosystems and aquatic ecosystems. Wetland, listed together with forest and ocean three global ecosystems, with an honor as "the Earth’s Kidney", maintains significant hydrological and ecological functions. Inland wetland within arid area is one of the important types in our country. However, compared with other wetland distribution areas, wetlands within inland barren regions and arid areas, particularly within barren plain regions are barely concerned by researches. As pointed out by The 32nd File from the State Council in 2007, development within the three counties of South Xinjiang is particularly supported to efficiently solve the poverty problem in south Xinjiang and to enhance the development in agriculture and industry. But the reality we face is an arid, desert, saline ecosystem with shortage of water resources, which is so fragile that it limits the social and economic development. As an obvious indicator and ecological factor, wetlands in arid areas have experienced the most active landscape changes by either natural impacts or human activities, such as wetland transfer, shrinkage or disappearance, all of which deeply show quality and volume of wetland, the living space of human being in arid regions. Whereas, wetlands in arid areas are mostly discrete in a small size, and are tightly related to ecological process of barren areas. Hetian oasis is located at the south edge of the Talimu Basin in Xinjiang, connected with the second largest desert on the earth, the Takelamagan desert in the north. The weather is extremely dry as warm continental desert region, and the ecosystem is vulnerable. In addition, the ground surface is mainly covered by gravel, sandstone, sand, which makes difficulty to develop wetlands and restoration from damage. Therefore, as one of important indicators of environmental change of oasis, it is of great significance to use remote sensing technologies to monitor the current status and dynamic changes of wetlands in time and accurately, which is very important in reasonable development, protection, sustainable development. Wetland systems are located at the boundary of water and land, and their characteristics are transitional. Land, wetland, and water are different parts of a continuous uniform system. Wetland has distinct seasonal changes and annual variations, as well as general heterogeneity in quality and homogeneity in spectrum or heterogeneity in spectrum and homogeneity in quality cause by its complexity and uncertainty. The major problem of wetland remote sensing study is to solve the problem to extract wetland information fast and accurate. This study utilize the third stage remote sensing data in the Hetian oasis area to explore wetland remote sensing technologies in arid areas in Xinjiang, and the study is shown as following,1) Based on wetland types, distribution and terrain in the study area, traditional wetland information extraction methods are firstly tested and evaluated. From spectrum analysis, traditional unsupervised and supervised classification methods lead to significant misclassification. For example, shallow river bed composed of gravel has high values is classified as barren areas.2) Decision tree is used for specific information extraction, which is built by image spectrum computation, eigen variable selection, PCA analysis, K-T transformation, NDVI variable.

Mots clés : Principle Component; Decision tree; Thermal infrared; Arid Zone Wetlands Index;

Présentation (CNKI)

Page publiée le 18 janvier 2018