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

A Method of Rain Removal Based on Image Restoration Technology

A technology of imagery, technology, applied in the field of deep learning and computer vision

Active Publication Date: 2022-06-03
TIANJIN UNIV
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, these methods do not make accurate use of the residual picture Like feature information to improve rain removal picture picture

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
  • A Method of Rain Removal Based on Image Restoration Technology
  • A Method of Rain Removal Based on Image Restoration Technology
  • A Method of Rain Removal Based on Image Restoration Technology

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0044] In order to make the objectives, technical solutions and advantages of the present invention clearer, the embodiments of the present invention are further described in detail below.

[0045] Considering that the above method does not accurately utilize the residual picture image feature information, and picture like repair techniques can target partially missing picture Like filling and repairing, the present invention proposes a novel picture Rain removal methods like restoration techniques. By building a threshold judgment module based on the existing rain removal network, the residual picture like convert to residual binary picture like and will go to rain picture Image and Residual Binary picture like feeding picture Like the repair module to repair high-quality rain removal picture Like, implement residuals picture Precise use of feature information. The rain removal method proposed by the present invention is beneficial to be based on picture Image obje...

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 invention discloses a method for removing rain based on image restoration technology. The method includes: constructing a rough rain removal network composed of three feature extraction modules and a residual image prediction module; constructing a threshold judgment module and an image repair module The composed image fine repair network; based on the distribution of the rainy image database, the rough rain removal network and the image fine repair network are trained, and the trained rough rain removal network and image fine repair network are combined into a rain removal model based on image repair technology; The input image is processed based on the deraining model. The present invention can accurately and effectively realize the utilization of the characteristic information of the residual image, can obviously improve the precision of deraining, and further promotes the actual scene application of the deraining algorithm.

Description

technical field [0001] The present invention relates to the field of deep learning and computer vision, in particular to a picture Rain removal methods like restoration techniques. Background technique [0002] Rain is the most common type of severe weather, in which rain lines of all directions, densities and sizes will block and blur the shot to varying degrees picture The content of the image directly leads to the deterioration of the subjective visual effect. In addition, poor quality picture image will also be significantly reduced based on picture performance on computer vision tasks like object classification, detection, and segmentation. Rain removal methods can be roughly divided into two categories: one is the traditional method based on dictionary learning, sparse coding, etc.; the other is the deep learning method based on multi-layer nonlinear convolution. Compared with the deep learning method, the rain removal obtained by the traditional method picture i...

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): G06T5/00G06T5/50G06N3/04
CPCG06T5/50G06T2207/20081G06T2207/20221G06N3/045G06T5/73Y02A90/10
Inventor 庞彦伟张雪岩
Owner TIANJIN UNIV