Active learning multi-source image fusion method under saliency driving
A technology of active learning and fusion methods, applied in the field of pattern recognition technology of support vector machines, to achieve good visual effects
Inactive Publication Date: 2020-02-21
BEIHANG UNIV
View PDF9 Cites 1 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0006] Aiming at the problem that the existing multi-source image fusion algorithm is only applicable to the fusion of two images, the present invention provides a multi-source image fusion method based on active learning, which can obtain each The probability map of the source image to obtain the fusion result
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
[0018] In order to better understand the technical solution of the present invention, the present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments. The concrete realization process of the present invention is as figure 1 As shown, the specific implementation details of each part are as follows:
[0019] Step 1: Perform saliency detection on multi-source images, extract their respective saliency regions, then weight the saliency regions and the relative gradient maps of each source image, obtain the feature map of each source, and carry out the feature map on each position Size comparison to obtain the salient point selection area of each source image.
[0020] Step 1-1: Perform Itti saliency detection on each source image, the specific flow chart is as follows figure 2 shown. For each source image, use Gaussian filtering to obtain I by 1 / 2 downsampling 1 , has been down-sampled by 1 / 2 until I get I 8 , s...
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 active learning multi-source image fusion method under saliency driving, and belongs to the technical field of digital image processing. The implementation process comprisesthe steps of 1) performing significance detection on a multi-source image, weighting a significance region and a relative gradient graph of each source image, obtaining a feature graph of each source, and performing size comparison on each position of the feature graph to obtain a significance point selection region of each source image; 2) partitioning each region according to spatial distribution, taking points from each sub-region of each source according to the position value of the feature map, and creating an eight-dimensional feature vector with a label; and 3) carrying out SVM classifier training on the obtained feature vectors, and iteratively training a classifier according to a strategy until the number of training samples is stable; and carrying out smoothing processing on theobtained probability graph, and then carrying out weighted fusion to obtain a final fused image. Through iterative training of the active learning classifier, the problem of fusion of more than two source images can be solved, and the method can be applied to the fields of target detection and the like.
Description
technical field [0001] An image fusion method oriented to multi-source images belongs to the field of digital image processing, and in particular relates to the image processing technology of saliency detection and the pattern recognition technology of support vector machine. Background technique [0002] For multi-source image fusion, especially the fusion of visible light and infrared images, because infrared images can detect targets emitting heat in special environments, it has a significant effect in reconnaissance, but general infrared images are limited by sensors. It is difficult to have a higher resolution, and the resolution of general visible light images is much higher than that of infrared images. Therefore, the visible light image and the infrared image can be fused, so that the fused image has a higher resolution and can detect some special targets at the same time. [0003] Most of the current multi-source image fusion methods are divided into two categories...
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 Applications(China)
IPC IPC(8): G06T5/50G06T7/00G06K9/62
CPCG06T5/50G06T7/0002G06T2207/10004G06T2207/10048G06T2207/20021G06T2207/20081G06F18/2411
Inventor 罗晓燕洪友勰申智琪于子淇
Owner BEIHANG UNIV