Remote sensing image scene classifying method on basis of multichannel layering orthogonal matching

Abstract

The invention discloses a remote sensing image scene classifying method on the basis of multichannel layering orthogonal matching and mainly solves the problem of low classification accuracy of the prior art. The remote sensing image scene classifying method includes steps of (1) setting up a training set and a test set for classifying remote sensing scene images; (2) intensively sampling images by adopting five different sliding windows to obtain image sampling points; (3) learning by a K-SVD algorithm dictionary; (4) sparsely encoding the image sampling points; (5) subjecting blocks of the images to maximum pooling; (6) setting up a second or third layer characteristic learning process for different sizes of image blocks obtained by the sliding windows; (7) using a pyramid model and maximum pooling to obtain image characteristic vectors; (8) classifying by a semi-supervised support vector machine. The remote sensing image scene classifying method sufficiently utilizes information of the images to set up different layers and paths of characteristic learning processes and can be used for scene detection and target identification of the remote sensing images.

Description

technical field [0001] The invention belongs to the technical field of image processing, relates to remote sensing image scene classification, and can be used for remote sensing image scene detection and image retrieval. Background technique [0002] With the rapid development of computer network technology and multimedia technology, remote sensing image scene classification has become a very important research field in image understanding, and has been widely used in image retrieval, computer vision and object recognition and other fields. Remote sensing image scene classification is a technology for automatic labeling of images based on image content. According to the type of features learned, it can be classified into two types: methods based on low-level features and methods based on middle-level features. Methods based on low-level features mainly include methods for classification based on color, texture, and shape; methods based on middle-level features achieve the pu...

AI Technical Summary

Problems solved by technology

Existing classical sparse coding methods such as LLC sparse coding, see references J.Wang, J.Yang, K.Yu, F.Lv, T.Huang, and Y.Gong.Locality-constrained linear coding for image classification.CVPR 2010; LSC sparse coding method, see reference Lingqiao Liu, Lei Wang, Xinwang Liu, "In defense of soft-assignment coding", in ICCV, 2011, pp.2486-2493. Both methods are based on Hard coding The improvement is obtained. Hard coding uses an atom closest to the local descriptor in the dictionary to reconstruct the descriptor. Such a tight constraint will lead to a large reconstruction error and the representation of the local descriptor is not flexible; LLC coding is It is to find the nearest K atoms to the descriptor and reconstruct the descriptor to make up for the defects of Hard coding; LSC is to find the K atoms closest to the descriptor, and then establish the relationship between the K atoms and the descriptor to be reconstructed The distance relationship, get sparse coding, lower time complexity than LLC

Images