Hyperspectral image dimension reduction method

Abstract

The invention discloses a hyperspectral image dimension reduction method, which comprises the following steps: firstly, dividing an original hyperspectral image into non-overlapped superpixels by using an over-segmentation method; next, as the pixel point in one super-pixel usually belongs to the same kind of objects, describing the spatial information by using an intra-class graph in the invention; and finally, introducing the intra-class graph based on the super-pixel level into an LGDE model as a regular item. In addition, in order to effectively capture the nonlinear characteristics of thehyperspectral image, the invention expands the linear LGDE into a kernel version. An original pixel point classification method (RAW), a principal component analysis (PCA), a linear discriminant analysis (LDA) method, a spectral space linear discriminant analysis (SSLDA) method, a local reservation projection (LPP) method, a collaborative graph-based discriminant analysis (CGDE) method, a sparsegraph-based discriminant analysis (SGDE) method, and a local graph-based discriminant analysis (LGDE) method are compared. Under the same experiment conditions, the classification result of the methodis more accurate.

Description

technical field [0001] The invention belongs to the technical field of image information processing, and relates to a hyperspectral image dimensionality reduction method. Background technique [0002] Remote sensing and airborne sensors are capable of measuring the spectrum of solar radiation reflected by the Earth's surface. Compared with multispectral sensors, hyperspectral sensors can provide richer spectral information. Therefore, hyperspectral sensors have become an important tool for the detection and classification of complex surface objects. However, some new challenges arise for hyperspectral image classification. The first challenge is the curse of dimensionality. Since labeling training samples is very time-consuming, we usually encounter hundreds of spectral channels but only a few training samples, which can easily lead to the Hughes phenomenon. The second challenge is the problem of information redundancy. Due to the dense sampling of spectral channels, hi...

AI Technical Summary

Problems solved by technology

However, some new challenges arise for hyperspectral image classification

The first challenge is the curse of dimensionality problem

Since labeling training samples is very time-consuming, we usually encounter hundreds of spectral channels but only a few training samples, which can easily lead to the Hughes phenomenon.

The second challenge is the problem of information redundancy

Therefore, intra-class structural information can be well learned, but inter-class information may not be fully captured

Images