Marine vortex recognition method based on MKL multi-feature fusion

Abstract

The invention discloses a marine vortex recognition method based on MKL multi-feature fusion. The marine vortex recognition method comprises the following steps: 1) carrying out data preprocessing ona data set based on a synthetic aperture radar image; 2) inputting the preprocessed synthetic aperture radar images into a feature extractor in batches, and extracting gray level co-occurrence matrixfeatures, Fourier descriptor features and Harris features; 3) constructing different types of kernel function sets, obtaining a training set of gray level co-occurrence matrix features, Fourier descriptor features and Harris features, and performing multi-feature fusion based on multi-kernel learning on the training set to obtain a data set; and 4) constructing a classifier model which is used forclassifying the data set. According to the invention, a plurality of feature fusion strategies are adopted, and a plurality of different types of features are applied to the identification of the marine vortex, so that the limitation of the data processing capability and the limitation of the conventional artificial visual and threshold setting method on the identification of the marine vortex inthe prior art are overcome.

Description

technical field [0001] The present invention relates to the field of remote sensing image recognition, in particular to a method for recognizing marine eddies based on MKL multi-feature fusion, which is used to synthesize aperture radar remote sensing images. Background technique [0002] The existing ocean eddy identification methods are: [0003] DiGiacomo et al. defined the visual features of eddies on SAR images by artificial recognition, and then analyzed the significant eddies on the southern coast of California from 1992 to 1998. [0004] Based on the visual characteristics of eddies on SAR images, Karimova et al. used manual interpretation to identify and analyze the eddies in the eastern Mediterranean, the Black Sea, and the Baltic Sea from 2009 to 2011, respectively. [0005] After smoothing the SST satellite image, Wang Chengyi represented and classified the obtained temperature contours with vortex features, and used the rough set theory to classify these closed...

AI Technical Summary

Problems solved by technology

[0017] 1) The early manual discrimination method was laborious and time-consuming, affected by the difference in the judgment of experts and supervisors, the uncertainty was large, and it was easy to produce unstatistical errors

[0018] 2) Traditional methods based on physical parameters and geometric features mostly rely on expert experience for manual feature design and single threshold setting. The size of the threshold changes continuously with different sea environment, and there are significant subjectivity and uncertainty.

[0019] 3) The recognition results based on image analysis methods mostly depend on feature selection, and the use of single feature recognition often cannot fully reflect the feature information of ocean eddies, which is prone to missed and wrong judgments

The traditional method based on physics and geometry is not universal, and the structural model is relatively complex, and it is impossible to identify the vortex with special shape and structure

[0021] 5) Although the method based on machine learning can learn and construct the characteristics of the vortex independently, the completeness of the training sample data limits the effectiveness of the extraction method. In the case of insufficient samples, it is not suitable to use the neural network method to train

Images