Study On Extraction Of Cage Culture Area From ZY-3 Satellite Images Based On Multi-feature

The rapid growth of aquaculture improves the development of fishery economy.However,the extensive development model has caused a series of negative effects on water environment.High-resolution remote sensing images are widely used in the field of object recognition and environmental monitoring because of their rich texture and spatial information.It has a great significance to extract aquaculture area using high-resolution remote sensing images for comprehensive analysis and evaluation of the aquaculture area development and ecological environment protection.In this paper,the ZY-3 satellite images were used as the data source,and the cage culture areas of Danjiangkou reservoir were selected as the extraction object.Based on the theory and method of automatic identification and extraction from remote sensing images,two methods of extracting the cage culture areas,i.e.multi-layer decomposition method and the multi-feature support vector machine(SVM),were proposed based on multi-feature analysis method.For the multi-layer decomposition method,the optimized normalized differential water index(ONDWI)and the texture mean value of the cage ratio index were used to eliminate other interfered objects.For the multi-feature SVM method,different eigenvectors are constructed by spectral and texture features to realized coarse extraction of the cage culture area,and the cage culture area was then refined using a custom extraction method.Compared with the maximum likelihood method and the minimum distance method,it was found that the multi-layer decomposition method and the multi-feature SVM method has a great potential in extracting the cage culture area from ZY-3 satellite images,the accuracy of cage culture extraction is above 81%.The result proved that the method developed in this paper,i.e.,the multi-feature analysis method,provides a practical way for the extraction of cage culture area from ZY-3 satellite images.Furthermore,it is of great significance for the cage culture area change monitoring in Danjiangkou reservoir area,and also for the ecological environment protection for the reservoir waters.