A semi-supervised classification method for hyperspectral remote sensing images based on the membership score of ground object categories
A technology of hyperspectral remote sensing and classification method, which is applied in the field of semi-supervised classification of hyperspectral remote sensing images, which can solve problems such as the degree of diversity aggravated by multiple scattering effects, the phenomenon of different objects with the same spectrum, and the same object with different spectra, etc., to achieve high-quality classification effect, high classification accuracy, good compatibility and robustness
Inactive Publication Date: 2018-01-12
FUDAN UNIV
View PDF5 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
In addition, factors such as low spatial resolution, heterogeneity of surface object distribution, and multiple scattering effects will aggravate the degree of diversity [2], which often leads to the phenomenon of different spectra of the same object or the phenomenon of different spectra of different objects, which makes classification difficult
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
[0061] Below, take actual remote sensing image data as an example to illustrate the specific embodiment of the present invention:
[0062] The semi-supervised classification method based on membership score in the present invention is denoted by SCAS, and the two modes that adopt SLIC and cube over-segmentation are denoted by SCAS1 and SCAS2 respectively.
[0063] real data experiment
[0064] We test the performance of the proposed algorithm using an actual hyperspectral remote sensing image dataset. The data set is the Indian Pines data set taken in 1992 by the Airborne Visible / Infrared Imaging Spectrometer (AVIRIS). The dataset contains 145×145 pixels, 220 bands, the wavelength range is 0.4-2.5μm, and the spectral resolution is 10nm. After removing bands with low SNR or water absorption, the remaining 186 bands were used for algorithm validation. figure 1 A pseudo-color map of the image is shown, along with a map of the true distribution of features. According to the fi...
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 belongs to the technical field of remote sensing image processing, and in particular relates to a semi-supervised classification method for hyperspectral remote sensing images based on the scoring of the membership degree of ground object categories. On the premise of over-segmentation, the present invention takes the membership score as the core, introduces the process of region growth, effectively combines spectral information and spatial information, and provides a new strategy for semi-supervised classification. Among them, the membership score is based on fuzzy theory, and simultaneously weighs three factors: spatial consistency, spectral variability and prior knowledge of hyperspectral images, and can obtain high-precision classification results and smooth classification identification maps. The present invention has good robustness to the parameters and the proportion of training samples in the total samples; the fuzzy scoring of the membership degree of the feature category efficiently utilizes prior knowledge, and only needs a very small amount of training samples to output high-quality classification results, and the classification accuracy is not sensitive to parameter changes; the invention has important application value in the classification of hyperspectral images.
Description
technical field [0001] The invention belongs to the technical field of remote sensing image processing, and in particular relates to a semi-supervised classification method for hyperspectral remote sensing images based on scoring of the membership degree of ground object categories. Background technique [0002] Remote sensing technology is a new comprehensive technology developed in the 1960s. It is closely related to science and technology such as space, electron optics, computer, and geography. It is one of the most powerful technical means for studying the earth's resources and environment. Hyperspectral remote sensing is a multi-dimensional information acquisition technology that combines imaging technology with spectral technology. The hyperspectral imager simultaneously detects the two-dimensional geometric space and one-dimensional spectral information of the target on dozens to hundreds of very narrow and continuous spectral segments of the electromagnetic spectrum,...
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
CPCG06F18/24133
Inventor 陈昭王斌
Owner FUDAN UNIV