Hyperspectral image classification method based on spatial feature iteration

Abstract

The invention discloses a hyperspectral image classification method based on spatial feature iteration. The hyperspectral image classification method comprises the steps of: calculating spectral signatures di of ground feature categories according to hyperspectral sample data; setting a category target set, a background end member set and a color constraint matrix; defining a classifier Tk capable of classifying multiple categories simultaneously by utilizing background end members U of a hyperspectral image, the constraint matrix C and an image autocorrelation inverse matrix R<-1>, extracting initial classification results of all category targets, extracting a spatial feature {Tk(iG)}of the classification result of each category target, and then feeding back and superposing the spatial features {Tk(iG)} into the hyperspectral image to be classified for carrying out fusion of the spatial features and inter-spectrum features; classifying multiple categories simultaneously by adopting an iteration mode until a set iteration stopping condition is reached; and marking classification results by using different colors. The hyperspectral image classification method effectively utilizes the method of fusing the spectral statistical features and the iteration spatial features to perform hyperspectral image category feature judgment, and gradually improves the accuracy of hyperspectral image classification.

Description

technical field [0001] The invention relates to a hyperspectral image classification method based on spatial feedback features, in particular to a hyperspectral image classification method based on spatial feature iteration. Background technique [0002] Hyperspectral images have the characteristics of high spectral resolution, and can detect types of ground objects that cannot be detected by multispectral images. They are more and more widely used in environmental monitoring, military fields, forestry and other fields. The goal of hyperspectral image classification is to classify and label each pixel in the hyperspectral image. The high spectral resolution and high spatial resolution of the hyperspectral image make it a huge advantage in the classification of ground objects. The accuracy of spectral information also makes interference and background parts have a certain impact on hyperspectral classification; at the same time, due to the characteristics of high-dimensional ...

AI Technical Summary

Problems solved by technology

However, this kind of method is sensitive to support vectors on the one hand, and the classification effect on some object categories will be poor; on the other hand, the kernel function is used to convert the data into a high-dimensional space for classification, which also causes memory consumption when the data sample becomes larger. Large, computationally intensive and other issues

At present, the hyperspectral image classification method based on interspectral-spatial features has become a current research hotspot. This type of method improves the accuracy of hyperspectral image classification by introducing spatial information features to assist spectral features. However, this method uses spatial features. It firstly extracts spatial features such as edge and shape texture of the image through morphological and filtering methods, and then uses support vector machines to perform one-to-one or one-to-many classification; on the one hand, it lacks the effective use of image spatial features, and on the other hand On the one hand, there is also a lack of fusion of spectral information and spatial information

Images