Hyperspectral and panchromatic image fusion method based on CNN and Laplace pyramid
A technology for hyperspectral images and panchromatic images, which is applied in image enhancement, image data processing, character and pattern recognition, etc., and can solve problems such as the difficulty of fully extracting details of panchromatic images, blurred images, and difficulty in establishing high-resolution hyperspectral images. Image, full-color image and other problems, to avoid the problem of gradient disappearance, reduce the propagation of forward and reverse information, and enrich the effect of spectral information
Pending Publication Date: 2021-04-16
XIDIAN UNIV
View PDF0 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0006] Through the above analysis, the existing problems and defects of the existing technology are: all CNN-based fusion methods are implemented by changing the network architecture of image super-resolution, and the CNN network is like a black box, which treats each band indiscriminately and cannot fully Extract spatial details of panchromatic images, which play an important role in image fusion
[0007] The difficulty in solving the above problems and defects is: due to the different spectral responses and the complexity of ground features, it is difficult for classic hyperspectral image fusion algorithms to establish Relationship
However, due to the limitations of the network architecture, the existing CNN-based methods are difficult to fully extract the detailed features of full-color images.
[0008] The significance of solving the above problems and defects is: in the process of hyperspectral image fusion, the extraction of detailed features is very important, which directly determines the quality of our final fusion results. If too much spatial detail is injected, it may lead to false images. On the contrary, if the extraction of spatial details is not enough, there may be problems such as image blur
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
[0071] In order to make the object, technical solution and advantages of the present invention more clear, the present invention will be further described in detail below in conjunction with the examples. It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention.
[0072] Aiming at the problems existing in the prior art, the present invention provides a fusion method of a hyperspectral image and a panchromatic image. The present invention will be described in detail below with reference to the accompanying drawings.
[0073] Such as figure 1 As shown, the fusion method of hyperspectral image and panchromatic image provided by the present invention comprises the following steps:
[0074] S101: Acquire an image data set, and preprocess the data;
[0075] S102: Decomposing a panchromatic image by using a Laplacian pyramid structure;
[0076] S103: Using the Dense block network to extr...
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 image processing, discloses a hyperspectral and panchromatic image fusion method based on a CNN (Convolutional Neural Network) and a Laplacian pyramid, and solves the problems of insufficient space detail information extraction and excessive spectral information loss of an existing hyperspectral and panchromatic image fusion method. The hyperspectral image and panchromatic image fusion method comprises the following steps: acquiring an image data set, and preprocessing data; carrying out panchromatic image decomposition by utilizing a Laplace pyramid structure; carrying out space detail extraction by utilizing a Dense block network; combining the low-spatial-resolution hyperspectral image with the extracted spatial detail features to perform image reconstruction; and carrying out model training by utilizing a loss function to obtain a final hyperspectral fusion image with high spatial resolution. According to the method, the spatial detail features can be fully extracted while the spectral information loss is reduced, the spatial information of the image is improved, and the high-quality hyperspectral fusion image can be obtained.
Description
technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to a hyperspectral and panchromatic image fusion method based on CNN and Laplacian pyramid. Background technique [0002] At present: hyperspectral technology represents the development frontier in the field of remote sensing, but due to the limitations of some physical conditions, a single hyperspectral sensor cannot obtain images with both high spatial resolution and rich spectral information. Panchromatic images imaged in a single electromagnetic spectrum have high spatial resolution but low spectral diversity. Hyperspectral images containing dozens or even hundreds of spectral bands are rich in spectral information, but have low spatial resolution. It can be seen that neither panchromatic images nor hyperspectral images alone can meet the requirements of target detection and change detection. Therefore, it is necessary to integrate the spatial information o...
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): G06T5/50G06K9/46G06K9/62G06N3/04G06F17/14
Inventor曲家慧徐云霜肖嵩董文倩李云松张同振
OwnerXIDIAN UNIV