Font Size: a A A

Object Oriented Remote Sensing Information Extraction Of Construction Waste In Wuhou District,Chengdu

Posted on:2021-05-26Degree:MasterType:Thesis
Country:ChinaCandidate:X Q ZhangFull Text:PDF
GTID:2381330647963111Subject:Cartography and Geographic Information System
Abstract/Summary:PDF Full Text Request
With the rapid development of urbanization in China,a large number of demolition,shantytown transformation,new projects etc produce a lot of construction waste under the current urban development planning in Wuhou District,Chengdu.Unreasonable disposal of construction waste will occupy large land resources and has a serious impact on the urban ecological environment,thus the disposal of construction waste become an urgent problem.Taking Wuhou District of Chengdu as the research object,the GF and Sentinel-2 images of Wuhou District as the research data,this paper uses the remote sensing image classification software e Cognition,and employs the object-oriented information extraction method to extract and analyze the construction waste in the research area.First,this paper applies the software,ENVI,for image preprocessing,then analyzes and extracts the remote sensing characteristics of construction waste and typical features in the research area.Second,we extract the construction waste and evaluate the accuracy of the extraction results using e Cognition software to calculate the optimal segmentation scale and multi-scale segmentation and employing the characteristics of construction waste to build the rule set of layered extraction.Finally,we derive and analyse the results of ground object extraction.Using object-oriented information extraction method to extract construction waste in the research area,we respectively calculate the coefficients of PA and Kappa of construction waste in GF-1,GF-2,Sentinel-2a and Sentinel-2b as following according to the accuracy evaluation: 64.11%,0.56;66.23%,0.63;60.02%,0.52;62.47%,0.54.It shows that the object-oriented information extraction method can effectively extract the construction waste,and can be used to extract the construction waste information in the research area.According to the analysis of the classified results,we find that 12.3% and 25.6% of the construction wastes in GF-2 and GF-1 are within 200 m from the water resource,while the values in sentinel-2a and sentinel-2b are 14.8% and 18.7% respectively.Moreover,Construction waste within 200 m away from vegetation accounted for 94.2%,98.0%,95.8% and 95.7% in GF-2 and GF-1,Sentinel-2a and Sentinel-2b,respectively.It is found that 57.1% and 61.8% of construction waste are converted into construction land,24.4% and 21.5% converted into vegetation,15.3% and 12.7% remained unchanged according to the analysis of the change of GF and sentinel-2 data in one year.The leached liquid generated by the construction wastes will seep into the soil and river,then pollute and damage the soil,water resource and vegetation,causing adverse impact on the ecological environment.In addition,the dust produced in the process of construction waste will reduce the air quality and even affect the health of residents.
Keywords/Search Tags:object-oriented information extraction, Construction Waste extraction, feature extraction, optimal segmentation scale
PDF Full Text Request
Related items