| Plant distribution of natural vegetation growing in different environment adaptation and selection results, greatly influenced by environmental factors, and it reflects the environmental effects on the survival and growth of vegetation. To get the vegetation information is the foundation research of natural. Coastal environment has multiple features such as vulnerable Ecotone areas, vegetation type, and vegetation changes rapidly, but there is serious confusion in its spectrum, and therefore fragile vegetation types of fine extracts has been remote sensing interpretation problems. It plays play an important role in environmental protection and recovery that mastering the fragile coastal area of vegetation cover information timely and accurately in monitoring the regional ecological environment, as well as for maintaining ecological stability and sustainable development in environmentally fragile areas is of great significance.The essay chose Kenli County, Shandong province the fragile coastal area as a study area. Landsat and MODIS data as the data resource, we used the method of the combination of object-oriented and vegetation phenology. Extract the vegetation information in Kenli County in 2004,2008 and 2013, and the three years of monitoring and analysis of vegetation variation of species change. Main contents and the results are as follows:(1) The study of method used to extract Vegetation Types. The essay used NDVI synthetic time series data on phenological information extracted in coastal areas. Study on the fitting method with the Logistic function. The images in 2004,2008 and 2013 were segmented by multiscale segmentation algorithm. Determine the division of different scale after many experiments. Using TM image in 2004 and 2008, through the test to determine segmentation scale of 250,200,150 and 200 respectively four segmentation scale, establish Level1, Level2, Level3 and Level4 four classification level. Use the data source for OLI 2013, segmentation scale of 300,250,200,250 and 200 divisions of five dimensions, respectively established Level 1, Level2 stores, Level3, Level4 and Level5 five classification levels. Combined with the phenological information and object-oriented method to extracting vegetation types in the study area. According to the study area and distribution of vegetation area of plaque size, to determine the different levels required for the extraction of vegetation category. The extraction of natural vegetation used the method of combination of phenology and object-oriented. Artificial vegetation caused by different cropping systems with different vegetation spectrum, we used the method of Object-oriented Neighbor and evaluate the accuracy of the results.(2) The research on spatial variation of vegetation in Kenli County. The classification change detection results of three years of treatment, the area of natural vegetation less 14.75% than 2004, artificial vegetation increased 4.65%, and from 2008 to 2013 natural vegetation area increased by 2.37%, artificial vegetation area decreased by 10.04%.2004 to 2013 reduced 12.39% of natural vegetation, in artificial vegetation area of 5.39%. The most obvious change among natural vegetation is Suaedas and reeds, Kenli County and the northern part of a significant reduction in space. In 2008, Suaeda has a significant reduction in the entire study area, near the mouth of the Yellow River that Suaeda area increased by a large range in2013.(3) Conclusions. The paper used the vegetation phenology and object-oriented method to 2013 Kenli OLI classify data, the confusion matrix method is adopted to improve the accuracy of the classification results, obtained overall classification accuracy of 80.75%, and Kappa coefficient of 0.79. With the 2013 OLI data, comparing the object-oriented classification results obtained total classification accuracy increased by 18.5%, the Kappa coefficient increased 0.21. The results show that the combination of vegetation phenological information extracted compared with phenological information extracted have been greatly improved. The result of change monitoring show that non-vegetation area was gradually increasing from 2004 to 2013 and the area of natural vegetation reduced from 2004 to 2008 and the artificial vegetation and vice versa. The increase of natural vegetation and artificial vegetation area is reduced in 2013, but overall, the total area of vegetation is reduced. The area of Suaeda and reeds has significant reduction among natural vegetation in the distribution area of inland Kenli County, Suaeda area showed an increasing trend in coastal areas. Artificial vegetation area has declined in inland areas except Lotus areas.The method of combine vegetation phenology information with object-oriented used to extract vegetation information and spatial variation of vegetation in different periods analyzing information for the development of ecological and environmental protection policies provided. |
PDF Full Text Request