Hyperspectral image segmentation method based on kernel method

A hyperspectral image and image segmentation technology, applied in image analysis, image enhancement, image data processing and other directions, can solve the problems of Haghes phenomenon and low image classification accuracy, achieve a wide range of use, realize hyperspectral image segmentation processing, large-scale The effect of applying value

Inactive Publication Date: 2017-01-25
HARBIN INST OF TECH
View PDF2 Cites 5 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The present invention proposes a hyperspectral image segmentation method based on a kernel method to solve the problems of low image classification accuracy and Haghes phenomenon in the prior art

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Hyperspectral image segmentation method based on kernel method
  • Hyperspectral image segmentation method based on kernel method
  • Hyperspectral image segmentation method based on kernel method

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0018] Specific implementation mode one: as figure 1 As shown, a hyperspectral image segmentation method based on the kernel method includes the following steps:

[0019] Step 1: Preprocessing the hyperspectral image data;

[0020] Step 2: clustering the preprocessed hyperspectral image data;

[0021] Step 3: Map the preprocessed hyperspectral image data and cluster centers to a high-dimensional space through a kernel function;

[0022] Step 4: Carry out image segmentation in the kernel function mapping space according to the image segmentation algorithm.

[0023] The hyperspectral image data is preprocessed, and the cluster center is found by clustering. Map the data and cluster centers to a high-dimensional space, and apply the kernel-mapped data to the graph cut function to obtain image segmentation results.

specific Embodiment approach 2

[0024] Embodiment 2: The difference between this embodiment and Embodiment 1 is that the preprocessing of the hyperspectral image data in the step 1 is as follows:

[0025] The KPCA method in the feature extraction method is used for dimension reduction processing, and the KPCA is a kernel-based principal component analysis.

[0026] In hyperspectral data analysis, measurement complexity is related to the number of bands and quantification accuracy. The more image bands and the higher the quantization accuracy, the greater the complexity of the data. There is an optimal data complexity that can optimize the classification accuracy. If the data dimension is high and the quantization accuracy is too high, the classification accuracy will decrease, which is the famous Haghes phenomenon. Dimensionality reduction is an important means to eliminate the Haghes phenomenon, so the preprocessing method adopted in the algorithm adopts the KPCA method in the feature extraction method fo...

specific Embodiment approach 3

[0028] Specific embodiment three: the difference between this embodiment and specific embodiment one or two is that: the specific process of clustering in the step two is:

[0029] Clustering is the process of dividing a collection of physical or abstract objects into classes, where each class is composed of similar objects. The collection of a group of data objects obtains a cluster generated by clustering. The objects in the same cluster are similar to each other, and the objects in different clusters are different from each other. In this algorithm, the K-means clustering method is used to cluster the data. .

[0030] Basic principle: The basic idea of ​​the K-means clustering method is to select K points from the data sample as the initial value of the cluster center, and move each cluster center successively during the iteration process until the cluster centers of two adjacent iterations do not change. , the clustering criterion function reaches the optimum.

[0031] ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention relates to a hyperspectral image segmentation method based on kernel method and is intended to solve the problems of the prior art that image classification precision is low and Haghes phenomena easily occurs. The method comprises the steps of first, preprocessing hyperspectral image data; second, clustering the preprocessed hyperspectral image data; third, mapping the preprocessed hyperspectral image data and a clustering center to a high-dimensional space through a kernel function; fourth, performing image segmentation on a kernel function mapping space according to image segmentation algorithm. The algorithm provides hyperspectral image segmentation, Haghes phenomena is eliminated by dimensional reduction, the method is suitable for visual analysis and mode recognition and allows hyperspectral image segmentation; the method is applicable to the field of image segmentation.

Description

technical field [0001] The invention relates to a hyperspectral image segmentation method based on a kernel method. Background technique [0002] Hyperspectral image is a combination of imaging technology and spectral technology. While obtaining two-dimensional scene information, it can also obtain image data of one-dimensional spectral information that characterizes its physical properties. Due to the rich information of hyperspectral images, how to improve the classification and recognition ability of spectral images has become a hot spot in current research and has great significance. Image segmentation is one of the most basic and important issues in low-level vision in the field of image processing and computer vision, and it is a prerequisite for image segmentation. Image segmentation is to divide the image into regions with different characteristics, and extract the target (or target) of interest, which provides a basis for subsequent analysis, understanding and clas...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06T7/11
CPCG06T2207/10036
Inventor 林连雷许永辉孙超
Owner HARBIN INST OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap