Unlock instant, AI-driven research and patent intelligence for your innovation.

Image fusion method based on convolutional neural network and saliency weight

A convolutional neural network and image fusion technology, applied in the field of image information fusion, to achieve the effect of improving detection efficiency and strong versatility

Active Publication Date: 2020-09-15
BEIJING INSTITUTE OF TECHNOLOGYGY
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The fused high-quality images of the present invention can provide strong support for subsequent target recognition, target detection, etc., and provide support for solving related engineering problems in the field of image fusion

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
  • Image fusion method based on convolutional neural network and saliency weight
  • Image fusion method based on convolutional neural network and saliency weight
  • Image fusion method based on convolutional neural network and saliency weight

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0063] In order to better illustrate the purpose and advantages of the present invention, the content of the present invention will be further described below with reference to the drawings and examples.

[0064] In order to verify the feasibility of the method, the source image is a visible light image and an infrared image, one for each, namely IR and VIS images. The neural network model net selects the VGG-19 network, and uses the four hidden layer name={relu1-1, relu2-1, relu3-1, relu4-1} in the network. The gray level of the image is L=256. The number of decomposition layers le=4. The final image fusion result uses multi-scale structural similarity MS_SSIM to objectively evaluate the selected 21 sets of infrared and visible images.

[0065] Such as figure 1 As shown, the image fusion method based on convolutional neural network and saliency weight disclosed in the present invention includes the following steps:

[0066] Step 1: Use guided filtering to decompose the base layer a...

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 image fusion method based on a convolutional neural network and a saliency weight disclosed by the invention belongs to the technical field of image information fusion. The invention realizes the decomposition of the base layer and the detail layer of the source image through guided filtering, and obtains the detail layer and the base layer of the source image; uses the saliency weight to fuse the decomposed base layer, and obtains a contrast-enhanced fused base layer image ;Multi-resolution singular value decomposition is performed on the source image detail layer, and the decomposed source image detail layer is subjected to multi-layer feature extraction and fusion of convolutional neural network to obtain the source image detail layer fusion containing fine details, and the reconstruction obtains high-quality fusion image. The high-quality fused image obtained by the invention has relatively high contrast information, contains the detail layer information of the source image, helps to highlight salient targets, and improves the detection efficiency of target recognition. In addition, according to the actual fusion requirements, the present invention changes the network structure in the method to achieve different fusion effects and has strong versatility.

Description

Technical field [0001] The invention relates to an image fusion method based on a convolutional neural network and a saliency weight, in particular to an image fusion method in a convolutional neural network, and belongs to the technical field of image information fusion. Background technique [0002] Multi-sensor data obtained by various sensors provides supplementary information through image fusion. Compared with images from a single sensor, image fusion produces good visualization and rich information. Therefore, it is widely used in many fields, such as remote sensing, pattern recognition, medical imaging and military. [0003] General image fusion methods are divided into four categories: (1) multi-scale decomposition; (2) sparse representation; (3) spatial domain transformation; and (4) hybrid transformation. Since the fusion method proposed by Laplace Pyramid, the typical image fusion method based on multi-scale decomposition theory has been applied to image fusion. In a...

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 Patents(China)
IPC IPC(8): G06T11/60G06N3/04
CPCG06T11/60G06N3/045
Inventor 郝群闫雷曹杰袁莉莉李国梁
Owner BEIJING INSTITUTE OF TECHNOLOGYGY