| Karst rocky desertification is a serious ecological problem in Southwest China, and vegetation restoration is the most important task of ecological restoration and rocky desertification control in Karst area. Chongqing, located in the upper reaches of the Yangtze River, and in the hinterland of the Three Gorges Reservoir area, is an important ecological barrier region, its environmental quality is directly related to the safety of the Three Gorges Dam and the Yangtze River industrial belt and the sustainable development of the southwest region. Northeast of Chongqing(Wuxi, Wushan and Fengjie) region is the key Karst Rocky Desertification control areas in Chongqing, this region has high mountains and steep slopes, complex terrain, serious vegetation degradation phenomenon. Therefore, it is of great realistic significance to master the vegetation distribution characteristics of Chongqing Zhongshan northeast region for the improvement of the environment and the control of rocky desertification.In this paper, based on the Landsat8 OLI remote sensing image, 30 m DEM data, combined with field survey data, the northeast region of Chongqing land use map data, remote sensing survey data, etc., Using the object oriented classification method in remote sensing software(eCognition 9.0, Envi5.0) platform to extract the region class. object oriented classification method integrated use of spectral remote sensing data, space, texture information to overcome the deficiency of traditional classification methods, improve the classification accuracy. Then the classification of the data using GIS software(ArcGIS), statistical software(Excel, SPSS) for statistical, mapping and analysis, and in the spatial layout of the summary and discussion, aimed at understanding the characteristics and laws of the region’s spatial and temporal distribution, provides a reference information for the rocky desertification control and restoration of the region, and provides basic data for the further study of energy in the region. The main conclusions of the study are as follows:(1)Research method: Northeast of Chongqing belongs to middle mountain area, terrain is complex and broken, consequently, object-oriented classification method was applied in eCognition 9 software to realize the extraction of feature information. Firstly, we deciding to use a large, medium and small(scale segmentation respectively 100, 60 and 40) three segmentation scale layered features extraction by the regional multi-scale segmentation experiment. Rivers, and difference larger vegetation and non vegetation area(not including rivers) are extracted on a larger scale; and in medium scale extraction vegetation zones of grass and shrub, farmland and woodland, non vegetation area of artificial construction and bare land; secondary class in forest and other small objects extraction in small scale. Then, the membership function method is used to establish the rule set and the extraction of the object information on different scales. Finally, the accuracy of the classification results was evaluated using field sampling and other data. It was found that the use of object-oriented classification method is more in line with the distribution patterns and characteristics of mountain terrain types.(2)Research results: From the point of classification accuracy, the overall classification accuracy reached 81.35%, which is basically come up to the requirements of mountain vegetation information extraction. The results showed that the land use types of woodland and cultivated land were the main land use types in the northeast region of Chongqing, The total area is 8872.22km2, accounting for the 96.50% of the study area. Along with the change of elevation and slope, the type of land cover has demonstrated obvious distribution change law.(1) There will be an enormous change at air temperature, soil, humidity and other physical factors with the increase of altitude. Therefore, plant species would be adapted to the specific environment. The classification map of ground objects and the analysis of the height data are found that the largest area accounted for 500 m to 1500 m of the area, accounting for 64.49% of the total area of the study area. Cultivated land and grass irrigation frequency is highest in elevation less than 1000 m, the altitude range is also the area where the distribution of frequent human activities. And at altitude greater than 1000 m region, arable land and grass irrigation frequency decrease quickly, and woodland frequency increased rapidly and the dominant. In all types of forest land, especially between the elevation of 500m-1000 m, horsetail pine and broad-leaved forest occupies the absolute advantage.(2) Different slope, which has resulted in the distribution soil, moisture, temperature and heat accumulation, will show a significant difference, also will affect the type and distribution of surface features on the slope. The analysis of the overlay analysis of classification map and slope data was employed and we found that: gentle slope, the slope and steep slope area is the main types and accounted for the total area of the study area, 21.34%, 30.41% and 25.31% respectively. From the point of view of various types of objects in different slope area distribution, Cultivated land, grass and forest land in the frequency and the change of the slope has a strong correlation.The overall trends are similar, performance for slope smaller than 25° were positively correlated, and the slope bigger than 25° negatively correlated. Embodied in the cultivated land and pine at level 1, 2, slope(0°-15°) appears the frequency highest, broad-leaved forest, mixed forest and grass irrigation times needle. The grass slope irrigation at level 2, 3(5°-25°) appears the frequency highest, then decreases gradually with the increase of slope. Forest land after slope bigger than 25° occupy obvious advantage, on behalf of vegetation species have pine and broad-leaved forest.(3)Comprehensive control measures : Since the policy of rocky desertification control has employed, Chongqing Northeast District, has been taking small watershed as a basic design and processing unit according to the distribution of various features, in accordance with the overall arrangement, the first easy to post difficult, classification of governance principles, the scientific selection of governance areas and measures. Specific mode of governance: 1.Ecological fragile areas dominated by vegetation measures, restoration and protection of existing vegetation; 2.Slope area in the implementation of the integrated management of sloping land engineering, improve land productivity; 3.industrial district, the implementation of orchard construction, increase farmer income and standard of living; 4.Animal husbandry area implementation of herbivorous animal husbandry engineering, changing the way of the production of animal husbandry industry. |
PDF Full Text Request