A multi-frame image deblurring system and method based on ER network
A multi-frame image and deblurring technology, applied in the field of computer vision, to achieve the effect of sharing security
Active Publication Date: 2022-06-03
安徽耕匠农业科技有限公司
View PDF7 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0004] Based on the above problems, the present invention provides a multi-frame image deblurring system and method based on ER network, which solves the problem that the image of atmospheric turbulence degradation has randomness, and there is obvious complementary information between multiple frames, and at the same time, there is no blur between multiple frames random discrepancies
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
[0065] The present invention will be further described below in conjunction with the accompanying drawings. Embodiments of the present invention include, but are not limited to, the following
[0070] The clearing processing module is used to clear the blurred target image uploaded by the user.
[0073] ERnet neural network model, including multi-frame complementary information extraction network, information refinement network and spatiotemporal attention
[0074] A degenerate feature extractor, comprising a first 2D convolutional layer, a first max pooling layer, and a first flatten layer, for output
[0075] A clear feature extractor, comprising a second 2D convolutional layer, a second max pooling layer, and a second flatten layer, for the output
[0079] Spatiotemporal attention mechanism, including 3D global draw pooling, 3D global max pooling, multiple fully connected layers, second residual
[0083] Atmospheric turbulence always carries out irregular motion, and the physic...
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 present invention relates to the field of computer vision, in particular to an ER network-based multi-frame image deblurring system, including an instruction acceptance module, an instruction generation module, an acquisition module, and a clear processing module; a multi-frame image deblurring method based on an ER network, Including step 1, construction of a degradation model, step 2, establishment of a simulation data set, step 3, generating the first degraded image and the first degraded image sequence for training, and step 4, using the ERnet neural network model to perform the first degraded image sequence Sort and evaluate the sorting results of the ERnet neural network model. Step 5. Add the output results of the multi-frame complementary information extraction network and the spatio-temporal attention mechanism as the input data of the information refinement network. Step 6. The user inputs the blurred target image , to generate a clear target image; it solves the problem that the atmospheric turbulence degradation image has randomness, and there is obvious complementary information between multiple frames, and there is also random difference between multiple frames.
Description
A multi-frame image deblurring system based on ER network and its method technical field The present invention relates to computer vision field, specifically refer to a kind of multi-frame image deblurring system based on ER network and its method. Background technique [0002] The restoration of degraded images is a key task in the field of computer vision, and it is a current hot research problem. The restoration method angle can be divided into non-blind restoration and blind restoration, non-blind restoration of the point spread function of the known image degradation, combined with the model Deconvolution of the blurred image and regularization rules, the complete blind restoration algorithm is where both the blur parameter and the noise parameter are unknown In the case of image restoration, the input is only the degraded image, and the non-blind restoration needs to predict the point spread function or point spread The function conforms to a specific prior dis...
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): G06T5/00G06V10/40G06V10/80G06V10/82G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06T2207/20081G06T2207/20084G06V10/40G06N3/045G06F18/253G06T5/73
Inventor 谢春芝高志升
Owner 安徽耕匠农业科技有限公司