A Study On Imagery Interpretation And Accuracy Test Of Urban Tree Canopy In Beijing Urban Area Based On ECognition

Urban forest establishment is one of the most significant approaches for combating the ecological and environment problems associated with the rapid urbanization.The urban tree canopy(UTC)is the best indicators for quantitively and qualitively assessing urban forest characteristics thus a primary measure for urban forest planning and management.With the availability of high-resolution imagery and digital classification techniques,assessing urban forest canopy cover is readily accessible to most urban forest managers.One method to handle highly variable pixel characteristics of urban forest canopy is through the use of Object-based image analysis(OBIA).OBIA allows for use of additional variables such as texture,shape,context,and other cognitive information provided by the image analyst to segment and classify image features,and thus,improve classifications.As part of this research we created UTC/LC classifications for a study area in Beijing,China,using OBIA techniques and 2002 aerial true-color imagery and 2013 WorldView satellite imagery with a resolution of 0.5 meter.The OA of 2002 classification results is 92.23% and 2013 is 96.02%.Kappa of 2002 classification results is 0.8613 and 2013 is 0.9231.We demonstrated that the OBIA approach can generate good UTC/LC classifications suitable for tree cover assessment in urban areas.Results demonstrate that:The study area had 95,984 hm2 of UTC(termed Existing UTC),comprising 39.53% of the total land area.8.69% of the area's land(21,098 hm2)could possibly be covered by UTC(termed Possible UTC-Vegetation),that is,there are grass or bare soil.The urban district generally had lower UTC,while the suburban district had higher UTC.Of the total land area,most UTC occurs on the land between 5th and 6th ring road in contrast to the land within 2nd ring road.The canopy cover derived from the 2002 aerial image of Portland was 19.46 %.These figures indicate an obvious increase in canopy over the 11 years.Similarly,the land between 5th and 6th ring road had larger degrees of change in UTC and the land within 2nd ring road had smaller.Within the Xishan limits,the Shijingshan District had the highest UTC,followed by Haidian District,Chaoyang District and Fengtai District which had similar UTC.Dongcheng District and Xicheng District had the lowest UTC.In the 11 years,Haidian District,Chaoyang District and Fengtai District had largest degrees of change in UTC,while Dongcheng District and Xicheng District had smallest.UTC in the First Greenbelt increased by 22.15 percent over the 11 years and in the Second Greenbelt increased by 23.97 percent.The reason for this change is the carrying out of large numbers of greening projects.Compared with green coverage,urban tree canopy cover has more outstanding advantages in ecosystem services assessment as it does not distinguish land ownership and emphasizes green space quality.