Hyperspectral Image Classification Method Based on Projective Structured Sparse Coding
A hyperspectral image, sparse coding technology, applied in the field of hyperspectral image classification, can solve the problems of limited classification accuracy, prone to errors, long processing time, etc., and achieve the effect of reducing cost and good classification
Active Publication Date: 2018-06-05
XIDIAN UNIV
View PDF4 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, this classification method is carried out separately for dimension reduction and classification, which is prone to errors and limits the classification accuracy to a certain extent.
At the same time, there are also problems such as long data processing time and insufficient use of neighborhood information.
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0038] The present invention will be described in further detail below in conjunction with the accompanying drawings.
[0039] refer to figure 1 , the concrete steps of the present invention are as follows:
[0040] Step 1, input hyperspectral image.
[0041] Input a hyperspectral image to be classified that contains k categories, set each pixel of the hyperspectral image as a sample, and each sample is represented by its band feature to form a feature vector, m is the number of bands of the hyperspectral image .
[0042] Step 2, determine the training samples and test samples.
[0043] (2a) Using the method of equal probability sampling, randomly select 10% of the samples in the labeled spectral vector of the hyperspectral image as the training samples, for any sample x in the training samples i , define a i 5×5 spatial window as the center, to obtain hyperspectral image training image block X i ;
[0044] (2b) Take the remaining 90% of the samples as the test samples ...
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
Login to View More Abstract
The invention discloses a hyperspectral classification method based on projection structure sparse coding, which mainly solves the problem that the prior art cannot effectively utilize hyperspectral image neighborhood information for classification. The implementation steps are: (1) read in the hyperspectral image data; (2) determine the training sample set and test sample set in the labeled spectral vector of the hyperspectral image; (3) solve the projection matrix according to the training sample set; (4) ) According to the projection matrix of the hyperspectral image training sample, calculate the sparse coefficient of the test sample set; (5) According to the projection matrix and the sparse coefficient of the test sample set, use the error discriminant function to judge the test sample, and obtain the test sample object category label. The invention has the advantages of high classification accuracy and low cost for processing high-dimensional data, and can be used to distinguish ground objects in hyperspectral images.
Description
technical field [0001] The invention belongs to the technical field of remote sensing image processing, and in particular relates to a hyperspectral image classification method, which can be used to distinguish ground objects in hyperspectral images. Background technique [0002] Hyperspectral image object classification is the main content of remote sensing technology processing, based on the following: the same type of pixel has consistency in spectral characteristics and spatial characteristics, and different object types have obvious differences in spectral characteristics and spatial characteristics. Hyperspectral image is a mass data of map integration generated by multi-spectral remote sensing imaging equipment, which contains spatial information of ground objects and non-negative spectral information. Each point in the image can be described by a high-level spectral vector composed of spectral information of many spectral bands, so hyperspectral data is characterized...
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
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/62
CPCG06F18/24143
Inventor 焦李成马文萍张风刘芳侯彪王爽杨淑媛
Owner XIDIAN UNIV