Study On Fast Segmentation Algorithm Of Island Boundary Based On Remote Sensing Image

China puts forward the strategic goal of building a maritime power in the period of "13th Five-Year".As an important carrier for the utilization and development of the ocean,the research on the rapid extraction of the island boundary,the rapid identification of the type of the land and the dynamic monitoring of its ecological environment,not only has important scientific value,but also has important strategic significance.Remote sensing technology is a far distance to Earth Observation technology,it can achieve large area simultaneous observation and quasi real-time dynamic monitoring.Based on large area coverage and quasi real-time acquisition of island remote sensing images,a fast extraction method of island boundaries is proposed.It provided a theoretical basis for the rapid census and mapping of the island,the contention of the sea power and the measurement of the land area.Image segmentation is an important method of image recognition,analysis and understanding.It has been widely applied to traditional image preprocessing.Remote sensing images have the characteristics of large area observation,real time,multi band and so on,when the traditional segmentation algorithm is applied to remote sensing image computation,it has many problems such as long computation time and low segmentation accuracy.Therefore,this paper analyzes the existing image segmentation algorithms,and combines the features of remote sensing images(large area coverage and multi band),and proposes a fast segmentation algorithm based on remote sensing images.The main contents of research are as follows:(1)In view of the multi band nature of remote sensing images,a rough segmentation method for island boundaries was proposed.In this part of research,a nonlinear optimization Otsu algorithm is proposed to achieve the rough segmentation of island boundaries based on remote sensing images.The main idea of the algorithm is to reduce the dimension of multi band remote sensing image by using PCA algorithm,reduce the number of bands and improve the computational efficiency on the guarantee the amount of remote sensing image information,and use the intra class variance of Sea Island and sea ground objects to replace the variance of the traditional Otsu,that is to make full use of the spectral characteristics of remote sensing image to optimize the termination condition of the algorithm,and improve the remote sensing image.The segmentation efficiency is achieved,and the island boundary was segmented(coarse segmentation).(2)In view of the large area coverage characteristics of remote sensing images,a fine segmentation method for island boundaries was proposed.In this part,two different rough segmentation algorithms are taken as input,and combined with the selective binary and Gaussian filtering regularized level set,an improved selective binary and Gaussian filtering regularized level set is proposed.The improved algorithm implements the fine segmentation of island boundaries based on remote sensing images.The main idea of the algorithm is that the rough segmentation result of the island boundary is taken as the initial evolution curve,and the initial evolution curve is close to the target boundary under the premise of ensuring the segmentation accuracy,which effectively solved the phenomenon of low calculation efficiency and over segmentation caused by the large coverage area of remote sensing images,and achieved the high precision optimization of the island boundary(fine segmentation).Taking the rapid extraction of the island boundary in the Taizhou sea area of Zhejiang as an example,it is found that the proposed method has obvious advantages over the traditional segmentation method in the calculation efficiency and the result precision.At the same time,taking the sea area of Zhoushan Archipelago as an example,the feasibility of rapid extraction of island boundary was discussed.