The Research On Extraction Of Urban Impervious Surface By LiDAR Combined With High Resolution Image

The large-scale extension of impervious surface resulted from urbanization,which turned(some)high permeable natural resources into impervious construction land,has caused a great impact/influence on urban ecological environment.The increase of urban impermeable surface with not only aggravate the pollution of water resources and vegetation reduction,but threaten cities with serious environmental problems.Urban impervious surface is not just an important indicator to measure the developing degree of urbanization,but an critical indicator to measure the changes of urban environment.It is of guiding significance for the sustainable development and planning of the city to extract the information of impervious surface accurately and effectively.How to classify the detailed features and extract the impervious surface from the complex urban environment is one of the focus and difficult issues in the field of remote sensing field.Taking a development area of Heng Yang City.Hunan province as a research area,this paper aims at extracting the impervious surface of urban areas.By combining mobile LiDAR data with high-resolution image data,applying object-oriented method and using the single image and corresponding classification method of multi-source data respectively,the research get a classification of extraction of impervious surface information of City HengYang.By carring on the comparative analysis,the research completes the high precise and detailed classification of urban impervious surface.To achieve fine-classification of urban land extraction provides a new thought and method for the remote sensing estimation of the high precise impervious surface.The innovation of the paper: first,complete detailed classification of city feature information by combining LIDAR with high resolution aerial images.Second,complete extraction of relevant features through impervious surface by using the new algorithm with the limitation of sole RGB images.Third,complete the discrimination of features which are easily confused by bombining the merits of LiDAR and high-resolution aerial images.The object segmentation scale and classification system proposed by this paper will provide an effective classification reference for such kind of image.The conclusions from this paper are as followings:(1)This paper introduces the object-oriented analysis technique,discusses the image segmentation algorithm and fuzzy mathematics classification method in detail,and get a more comprehensive overview of the theory and method of multi-scale segmentation algorithm.The selected methods of segmentation parameters were discussed in detail as well.In addition,this paper focuses on the theory of fuzzy classification,and on the basis of which clarifies the classification rules and systems.(2)The supervised classification process and system for R.G.B High-resolution aerial images of the research area were constructed.The Multi-scale segmentation is the core and classicfication premise of the object-oriented technology.Therefore,the paper segments the images into multi-scale before all classification steps,and finds that the segmentation parameters are applicable to various urban features.Finally,according to the rules and method of supervised classification,detailed classification of features and urban impervious sfuface are obtained.(3)The paper establishes a set of technical process,classification system and rules for information extraction of urban impervious furface.It extracts the urban impermeable surface through integrating LiDAR data with Aerial images acquired by two kinds of data sources from one equipment.Besides,the paper sets up a method which is available for extracting the urban impervious surface with fuzzy classification method.Moreover,the paper evaluates and analyzes the result by applying the best classification and verification method of confusion Matrix Accuracy.The result shows that this method can obtain some effective information of impervious surface of complicated urban features.(4)Multi-source data cooperation method can improve the classification accuracy of images.The urban impervious surface can be extract by means of the single aerial image and the combination of image and LiDAR data.Comparing the accuracy classification evaluation of these two means,the paper finds that the multi-source data can be complemented by the advantages of data when extracting impervise surface,so that the classification rusults are optimized.At the same timeby applying the combination of these two different characteristics.Some detailed and available classifications and references for the features of ill-separability,such as bare land which were proposed,therefore,The paper acquires some more precise classification results.