Study On High Resolution Remote Sensing Extraction And Classification Of Zhanjiang Offshore Aquaculture Area

With the rapid development of satellite remote sensing technology,remote sensing technology has a wider range of observations,short revisit cycle,and the advantages of terrain information rich increasingly highlight.It was widely used in agriculture,fishery production activities,and gradually replaced manual inspection type of traditional agriculture,fishery resources investigation form.Remote sensing become the agricultural,fisheries resources investigation and monitoring of important technical means.Zhanjiang city,located in the east of Leizhou peninsula,has rich fishery resources and a long history of fishing and aquaculture.It is one of the main aquaculture areas in Guangdong province.Remote sensing monitoring technology in aquaculture areas in China is mainly used in Fujian province,Jiangsu province,Shandong province and other regions at present,which can significantly improve the economic and social benefits of the region.However,there are few reports on remote sensing monitoring research in Zhanjiang aquaculture areas.Therefore,it is of great practical significance to carry out the monitoring research of Zhanjiang coastal aquaculture area based on remote sensing technology and improve the application level of remote sensing technology in the field of aquaculture in Guangdong province.The paper presents a method of multi-feature analysis,which is based on the feature of ground objects and combines the threshold method and object oriented analysis method,and the offshore raft aquaculture area and the aquaculture pond area of Zhanjiang city,Guangdong province were taken as the research object,and the domestic high-resolution GF-1 PMS image and ZY-3 01 A image were used as the data sources.The basic idea of multi-feature analysis is to construct feature index by using multi-feature index method,distinguish breeding area from seawater and other features,and select feature sensitive band to form feature set.Then the improved constrained energy minimization algorithm is used to enhance the target ground objects in the breeding area,weaken the background spectral information,and enhance the remote sensing image.Then,Otsu algorithm is used to calculate theoptimal threshold value of ground object extraction,and the difference between the result of Otsu extraction and the threshold segmentation of the single-band image is calculated to extract the target ground object preliminarily.Finally,according to the characteristics of the distracters in raft-type aquaculture area and aquaculture pond area,the distracters in raft-type aquaculture area and aquaculture pond area after preliminary extraction were customized and removed by using the gray co-existence texture matrix and the object-oriented method of geometric shape threshold segmentation,so as to realize the accurate extraction of aquaculture area.In this paper,the extraction accuracy of multi-feature analysis method is evaluated quantitatively,and the neural network method,support vector machine method and maximum likelihood method of traditional supervised classification method are used for comparative analysis.The results show that even in the breeding areas where the phenomena of "homologous foreign body" and "homologous alien spectrum" are obvious,the algorithm can still overcome the background ground object interference effectively,and obtain the extraction results of raft breeding area and pond area with high accuracy.Compared with supervised classification,the overall accuracy of the algorithm is more than 96%,and the Kappa coefficient is more than 0.93.The result of precision evaluation shows that the multi-feature analysis method proposed in this paper can accurately extract the aquaculture features in the coastal waters of Zhanjiang bay and can well adapt to different domestic high-resolution remote sensing images.The results of this study can provide effective method support and decision-making basis for relevant fishery management departments,and help to improve the capability of scientific planning and management of coastal aquaculture.