Study On Vegetation Change Detection Based On Spatio-temporal Correlation

Vegetation Change Detection(VCD)is a hot topic of remote sensing research field.On the basis of existing related research achievements,this thesis presented a change detection method based on spatio-temporal autocorrelation to deal with the limitations of current methods proposed in literatures in accuracy,fault tolerance of image data and registration,and controllability in manual handling processes.The empirical study has also been completed by taking different types of land cover changes in Changting County.Fujian Province as an example,which may promote the development of the field of VCD,and provide more scientific and reliable technical support for soil erosion assessment at the same time.It has important academic value and clear prospects.The main contents and results are as follows:(1)It proposed a VCD method based on spatio-temporal autocorrelation.The method made full use of the spatial neighborhood relations to grasp the overall region changes and the detail fature changes at the same time,and reduced the requirements of the quality of remote sensing image and the accuracy of registration,which improves the accuracy,efficiency,universality and robustness of detection to a certain extent,and would bettermeet the application needs for this technology.(2)It optimized the threshold used in VCD and the process of noise filtering.The former was completed with objective function automatically which ensures the objectivity of the results.The latter was achieved with spatial autocorrelation which improved the accuracy of change detection.(3)It made a comparison among the proposed method and other existing common change detection methods in terms of differences in accuracy.The result showed that the proposed method is better than the current representative VCD methods which based on pixels.(4)It acquired the result of vegetation change detection of Changting County by utilizing the method proposed in the thesis.Analysis based on the linear trend method indicated that,between 2009-2013,the varying intensity of artificial vegetation was bigger than that of natural vegetation,both of them appeared a positive trend,and compared with the northern,western and southern region,the eastern and central region was better with respect to vegetation recovery.(5)It completed the succession analysis on land cover change of each town in Changting County.The result showed that there was generally positive vegetation succession in Changting County.The positive succession trend was significant in central and eastern regions,while there were of weak positive,light degradation,moderately degraded and reverse succession trend in low mountains and hills of northern,southern and western regions which are still the key areas for erosion control in future.