Font Size: a A A

Segmentation Scale Optimization Of High Resolution Remote Sensing Image In Coastal Zone

Posted on:2017-05-11Degree:MasterType:Thesis
Country:ChinaCandidate:Y Y HuFull Text:PDF
GTID:2392330485467838Subject:Photogrammetry and Remote Sensing
Abstract/Summary:PDF Full Text Request
The remote sensing technology is an effective tool to obtain coastal typical objects with high spatial resolution.Therefore,it has great theoretical value to extract high-precision land cover/land use information for the sustainable development in coastal zone,and is of practical significance for reasonable allocation and comprehensive exploitation of marine resources.This paper aims at existing problems of object-based image analysis application to monitoring resources and environment in coastal areas,and has developed series of researches which contains multi-sclae image segmentation as well as optimal image object detection in coastal zone.The specific contents and conclusions are as follows:In this paper,the high spatial resolution remote sensing images from IKONOS-2 and WorldView-2 satellite are used in the segmentation research,and there are four research areas that come from Qingdao and Wenzhou province.A new segmentation method for high spatial resolution remote sensing image which combines the global edge and region information is developed from a new scheme to monitor the best conditions for each growing object to obtain the corresponding meaningful image object during multiscale analysis.The edge completeness curve used for getting a meaningful image object 1s the accumulation of those complete values generated in multiscale segmentation,and the selected scales lying in maximal completeness coordinates in the curve of an object are candidate segmentation scales for the object.The algorithm is carefully designed in consideration of two points:it adopts the global edge statistical parameter beside the region contour rather than the local edge information,which alleviates the effect of false edges.Furthermore,the extraction of image objects comes from the automatic curve analysis without any human involvement.The assessment for the segmentation results shows its merit in coastal remote sensing.The results demonstrate the ability of the proposed method to detect regions that have weak boundaries in images at their precise shapes.Although accurate edge positions near the important region boundaries are still necessary,the interrupt feature of an edge has less impact on the method.In addition,the results can reflect the spatial distribution of dominant ground objects with substantial differences in shape forms,which are hardly obtained from traditional one-scale segmentation.Generally,our optimal image object detection method can get a high accuracy in the remote sensing investigation of resource environment of coastal zone,and has technological innovation as well as practical application value in the field of high spatial resolution remote sensing image.
Keywords/Search Tags:high spatial resolution remote sensing, object-based image analysis, multi-scale segmentation, optimal image object detection, edge detection
PDF Full Text Request
Related items