Extracting Marine Floating Raft Region Based On Fusion Of Optical Image And Radar Image

China which has abundant shallow water and beach resources is the only country in the world where the production of aquaculture is higher than the catch of aquatic products,and the mariculture industry has been developed and flourished.However,in recent years,water pollution and eutrophication endanger the normal development of aquaculture industry;On the other hand,with the development of the coastal shipping and aquaculture industry,the contradiction between the growing number of transport vessels and the expanding aquaculture zones is becoming increasingly prominent,which make the regulatory authorities pay more attention to the regulation and management of the mariculture industry and aquaculture monitoring has become a hot research direction.The traditional floating raft aquaculture extraction is artificially using GPS to carry out field measurement,which is time-consuming and laborious,and the precision of monitoring results is not high.Remote sensing technology which has the characteristics of large scale,strong macrocosm and high timeliness has unique advantages in the extraction and monitoring of coastal aquaculture area.At present,more algorithms have been developed for the extraction of floating raft aquaculture by remote sensing image at home and abroad.However,on the technical level,conventional optical remote sensing or microwave remote sensing has its own limitations in time resolution and spectral information utilization.Therefore,in order to extract floating raft culture area more accurately,this paper proposes a method of applying the image fusion algorithm to extract floating raft culture area,and explores the fusion method which is most suitable for floating raft culture area extraction through experiments.Firstly,the high resolution SAR image and optical image are fusion by four fusion methods,whch integrates the advantages of optical and SAR remote sensing,and the fusion image is evaluated qualitatively and quantitatively to is obtain the method corresponding to the best fusion image.In order to verify the fusion method which got the best fusion results is the most suitable method for extraction,in this paper,the extraction work is carried out for floating raft culture areas in offshore waters under cloudy and foggy environment.It is verified that the fusion method which is most suitable for the extraction of floating raft culture area near the coast is the best fusion method obtained above.The method and results can provide important support for extraction and dynamic monitoring of floating raft culture area and provide data basis for each research.The main lines of work of this paper are as follows:1.At present,remote sensing images used to extract marine floating raft culture areas can be divided into visible light images,SAR images and unmanned aerial vehicles images according to different sensors and platforms.The main methods used for three kinds of image extraction in mariculture areas at home and abroad include artificial visual interpretation,index method and object-oriented classification.We analyzed the respective principles of the three methods and explained their advantages and disadvantages and their application in the extraction of floating raft culture in seawater.We also introduced the concrete implementation cases of the mainstream methods at home and abroad.2.As can be seen from domestic and foreign research examples,most of the remote sensing data used in the current research on the extraction of floating raft areas are single optical images or single SAR images,and there is no research on the extraction of the advantages of the two images.In this paper,the image fusion algorithm is used to extract floating raft region,and the commonly used fusion algorithm is introduced,and the fusion method which is most suitable for floating raft breeding area extraction is explored by experiments.Different fusion methods are used to fuse the optical images and SAR images from three different latitudes of Changhai County,Shandong Province and Sanduao,Fujian Province,including HIS transform fusion,Brovey fusion,Gram-Schmidt transform and K-L transform.The resolution of the fused image is consistent with the high-resolution SAR image.It also inherits the rich spectral features of optical images and transforms them into high resolution images with spectral information.Then the fusion image is evaluated qualitatively and quantitatively,and the corresponding method of the image with the best fusion effect is obtained.3.To explore the most suitable fusion method for extracting floating raft culture area,four different fusion results were classified by object-oriented classification to extract seawater floating raft culture area,and then qualitative and quantitative evaluation was carried out.According to the parameters of quantitative evaluation,such as the overall precision and KAPPA coefficient,the results of extraction of different floating raft culture areas were evaluated qualitatively and quantitatively.Finally,the fusion method which is most suitable for the extraction of floating raft culture area is obtained and it is verified that the best method of fusion is the most suitable method for extraction.