Study On Urban Green Construction And Management Based On 3S Technology

Urban green space, a major carrier of urban ecological system, is one of important indices to evaluate city civilization and capability of Sustainable development. As a producer and supervisor of urban green, city green management plays an essential roll for green source investigation and supervision, and takes responsibility of urban green plan and construction. In the dissertation, Based on the analysis of urban green definition and services, four issues are discussed in detail, which are two green information detection methods by remote sensing images, development of GPS(Global Position System) field survey tool for green information, application of GIS(Geographic information system) in green analysis and planning, implement of vegetation investigation and management information system. The main research conclusions of this paper are summarized as follows:(1) Object-oriented green cover information extractionAlthough the concept of object-oriented has been presented to remote sensing area in 1970's, pixel-based classification has been applied in information extraction from remote sensing images. Now, with the rapid develop of remote sensing technology, image analysis method based on pixel spectral property can't satisfy the demand of high spatial resolution image classification any more. Meanwhile, object's context, geometry and topology character joined with spectral character in classification is the most important advantage for object-oriented classification technology. In this paper, based on deep analysis in its theory and method, green information like arbor, shrub and grass in Yanzhong Park of Shanghai is detected by object-oriented classification which provides a method of fuzzy classification by object property set based on image segment. Compared to conventional pixel-based supervised classification, the total precision of new method is much higher.(2) Research and implement of semi-automated individual tree crown detection algorithm based on transect sample curve fittingAs an import part of urban ecological system, trees can provide remarkable ecological profits and its crown which provides a place to photosynthesis and energy production is a notable evaluate index for urban green ecological profit. The ordinary methods to extract crown area are field survey and manual interpret in remote sensing image. In this paper, A semi-automated individual tree crown detection algorithm is developed which has two main steps: firstly, several spectral characteristic transect samples are fitted by high-order curve approximation method; secondly, each curve inflection point is found to connect to tree contour. Two different types of tree areas are selected to detect crown area, speed can increase several times to manual interpret. In the first image street trees areas have been extracted where few of tree crowns were overlain, its detection precision is 87.82%; In the other image individual tree crowns have been extracted where more tree crowns are overlain, its detection precision is 84.42% a bit lower than the former.(3) Study on building shadow area correction for remote sensing interpretation dataBecause of sensor's precision, captured period and building shadow area, green interpreter data must be corrected to reduce and remove three system errors above. This dissertation focuses on high building displacement correct in remote sensing images. Because the shadow is far away from the image principal point and is concerned with three facts, distance between building and principal point, building shape and height, correct model can be different for various images. Based on analysis of imaging characters of two common remote sensing images, defilade area formulas are deduced, and the hidden green area can be calculated by correcting green cover rate or experience model. Taking Changqiao No.5 New Estate in Shanghai as research area, green areas of building shadow both in color infrared aerial photo and QuickBird image have been calculated and green cover rate are corrected, the result shows that the accuracy can be increased by 4-5 percent.(4) Prescription study on urban green resource investigationUrban green investigation criterion by remote sensing technology and GPS field survey criterion are built up according to real product prescription and flow. For RS criterion, investigation content, technique index, product flow, quality control, accuracy demand, result format and so on are detailed discussed. For GPS criterion, GPS selection method is defined by instrument capability and survey accuracy. Besides, collection method, product demand, project transform and position precision by position accuracy and relative position precision are also included in this criterion.(5) Urban ecological network construction based on shortest weight pathBased on analysis of conventional urban ecological network creating method, a shortest weight path method is proposed. Taking an area in the Inner Elevated Ring Road of Shanghai as study area, green space is divided to four types which are large patch, normal-large patch, normal patch and small patch, and passages composed by road, river and green belt are assigned to different values for their ecological function and green degree. Consequently, four network schemas are built up by different patch types and evaluated by network structure analysis. An optimized schema is founded by overlap analysis.(6) Survey and management information system for urban greenAn urban green survey and management information system is developed and implemented. By integrating green space detection tools in remote sensing images and analysis and evaluate fuctions, the system can be used for urban green's investigation, monitoring and management, green project's check and accept, green index computation and green space plan.