Object spectral feature extraction method based on spectral space mapping and purification
A spectral feature and spectral space technology, applied in the direction of instruments, character and pattern recognition, computer components, etc., can solve the problems of retaining the original spectral features of images, spectral information distortion, loss, etc.
Active Publication Date: 2017-11-24
三亚中科遥感研究所
View PDF8 Cites 1 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, in the process of image fusion, distortion and loss of spectral information often occur, and the original spectral features of the image are not preserved to the greatest extent. The quality of image fusion results directly determines the effect of feature extraction and the accuracy of final feature application.
At the same time, the feature extraction method based on image fusion is suitable for the fusion between two data source images and the object spectral feature extraction, and there are certain limitations in the object spectral feature extraction of three or more different resolution data sources
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
[0025] 1. attached figure 2 The SPOT-5 high-resolution remote sensing image used for object spectral feature extraction, sub-image a is a panchromatic image, 1 band, and the image size is 1200×1200 pixels; sub-image b is a segmented image of the panchromatic image, image The size is 1200×1200 pixels; the sub-image c is the multispectral image corresponding to the panchromatic image, with 4 bands, and the combination of 321 corresponding RGB bands is used, and the image size is 300×300 pixels.
[0026] 2. The resolution ratio between the high-resolution panchromatic image and the multispectral image is 1:4. According to the spatial correspondence between pixels, the spectral features of the multispectral pixel are directly mapped to each image of the panchromatic image. On each pixel, the multispectral initial mapping result S of the panchromatic image is formed, with image 3 The initial mapping of the pixels of the panchromatic image to the multispectral feature space is sh...
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 an object spectral feature extraction method based on spectral space mapping and purification. According to the method, first, an initial spectral feature space S of a full-color image is obtained according to spatial relation mapping between pixels of the full-color image and pixels of a multispectral image; second, multispectral pixel object purity is defined based on a segmentation result O of the full-color image, and a multispectral pixel object purity graph is obtained; and last, an object purity threshold value and a mark image B with the same size as the full-color image are set, the corresponding position with the object purity higher than the threshold value of B is marked as 1, the corresponding position with the object purity lower than the threshold value of B is marked as 0, a mean value of initial mapping spectral features of the pixels marked as 1 in an object is subjected to statistical analysis by combining S, B and O, and spectral features of the object are obtained. The method is suitable for spectral feature extraction of the object between the full-color image and the multispectral image of a single data source of a high-resolution remote sensing image or spectral feature extraction of the object between multi-resolution data sources in the object analysis process, and can be widely applied to object-oriented special target recognition, land cover classification, change detection and other fields.
Description
technical field [0001] The invention is an object spectral feature extraction method based on spectral space mapping and purification, which is suitable for object-oriented analysis between a single data source of high-resolution remote sensing images, panchromatic and multi-spectral images, or between multiple data sources with different resolutions Object spectral feature extraction. The invention can be widely used in the fields of object-oriented thematic target recognition, land cover classification and change detection, and the like. Background technique [0002] With the rapid development of earth observation technology, a series of high-resolution remote sensing satellites have been launched one after another, which has greatly improved the ability to obtain high-resolution earth observation data, and humans have entered the field of multi-source high-resolution earth observation data acquisition. New Era. Traditional remote sensing image information extraction is ...
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 IPC IPC(8): G06K9/46G06K9/34
CPCG06V10/267G06V10/56
Inventor 何国金王桂周袁继颖张兆明王猛猛彭燕程博白雅卿张晓美刘慧婵
Owner 三亚中科遥感研究所