Construction and merging classification method for high spectrum data multi-characteristic space

Abstract

A construction and merging classification method for high spectrum data multi-characteristic space comprises the following steps: (1) obtaining high spectrum reflection rate data and constructing high spectrum initial characteristic space; (2) utilizing a real-dimension analysis method to determine the category number of objects to be classified; (3) utilizing a semi-automatic image unit trainingsample selection method to obtain training samples to be classified; (4) determining a single category separability measurement criterion according to a largest separability principle; (5) utilizing an optimizing algorithm to obtain weight optimizing characteristic space according to the separability measurement criterion determined in the step (4); (6) performing single category optimizing linear transformation on the weight optimizing characteristic space to obtain linear transformation characteristic space specific to single category optimization; (7) performing classification on the linear transformation characteristic space to respectively obtain a classification result specific to the single classification optimization; and (8) designing merging rules, merging the classification result of the single category optimization and obtaining an accurate merging classification result.

Description

technical field [0001] The invention relates to a multi-feature space construction and fusion classification method for hyperspectral data, which belongs to the field of hyperspectral data processing methods and application technologies, and is suitable for research on theoretical methods and application technologies for high-precision classification of hyperspectral data. Background technique [0002] Hyperspectral remote sensing can obtain the spectral information of ground objects in a near-continuous narrow spectral band, characterize the characteristics of ground objects in more detail, and provide richer information for the distinction of ground objects. The research on the classification technology of hyperspectral data has always been one of the main research directions of hyperspectral remote sensing applications. Outstanding achievements have been made in various fields such as biomedicine. [0003] Different from broadband remote sensing, hyperspectral data has t...

AI Technical Summary

Problems solved by technology

[0003] Different from broadband remote sensing, hyperspectral data has the characteristics of multi-band, nearly continuous spectrum, large data volume, and large information redundancy, which brings great challenges to the subsequent processing and analysis of hyperspectral data: The processing of multi-dimensional feature space requires a lot of calculations; the hyperspectral data bands are highly correlated and redundant, and its core information is actually contained in relatively low-dimensional data; as the feature dimension increases, the data in high-dimensional The space shows different characteristics from the low-dimensional space, so the processing method of high-dimensional features is different from that of low-dimensional features, and the problem is more complicated; the number of training samples required by a supervised classifier is a function of dimensionality, and limited training samples restrict existing With the application of classification technology, when the number of training samples is limited, the classification performance will not continue to improve with the increase of the dimension, but there is an optimal dimension, directly analyzing and processing the hyperspectral data with limited training samples, the effect is often not good. good

Images