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

Single Image Dehazing Method Based on Prior Knowledge Guided Conditional Generative Adversarial Network

A technology of condition generation and prior knowledge, applied in the field of image processing, can solve problems such as strong data dependence and lack of real scene comparison, and achieve high fidelity, easy promotion and use, and novel ideas

Active Publication Date: 2020-04-24
ROCKET FORCE UNIV OF ENG
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, this method is very dependent on data due to the lack of comparison of real scenes.

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
  • Single Image Dehazing Method Based on Prior Knowledge Guided Conditional Generative Adversarial Network
  • Single Image Dehazing Method Based on Prior Knowledge Guided Conditional Generative Adversarial Network
  • Single Image Dehazing Method Based on Prior Knowledge Guided Conditional Generative Adversarial Network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0037] Such as figure 1 As shown, the single image defogging method based on the prior knowledge-guided conditional generation confrontation network of the present invention includes the following steps:

[0038] Step 1. Establish a fog image training set: use the image data set of known depth to synthesize a group of foggy image training sets according to the atmospheric scattering model, effectively expanding the image data volume of the foggy image training set;

[0039] In this embodiment, the image data set of known depth includes the NYU image data set.

[0040] Step 2. Preliminary dehazing of a single random foggy image: Randomly extract a foggy image from the fogged image training set in step 1, according to the formula Preliminary Dehazed Image Obtained by Prior Knowledge Among them, I h is a single random foggy image, A is the global atmospheric illumination, and T is the medium perspective;

[0041] In this embodiment, in step 2, the global atmospheric illumin...

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 defogging a single image based on a priori knowledge-guided conditional generation confrontation network, comprising steps: 1. Establishing a training set of fogged images; 2. Preliminary defogging of a single random foggy image; 3. Dehazing training of the preliminary dehazing image; 4. Calculating the true and false values ​​of the defogging training image of the reference truth image and the preliminary dehazing image; 5. Calculating the image loss objective function; 6. Updating the weight parameter set; 7. Calling For a new single random foggy image, repeat steps 2 to 6 until the true and false values ​​reach the set value; 8. Defog the single actual foggy image. The present invention uses prior knowledge to guide the encoding network to generate fog-free results, utilizes some useful information obtained from prior knowledge, and utilizes the feature modeling ability of deep neural network to make up for the lack of prior knowledge. The atmospheric scattering model is shown in the deep neural network, but it is regarded as the conditional generation of the image, and the dehazing effect is good.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to a method for defogging a single image based on a priori knowledge-guided conditional generative confrontation network. Background technique [0002] Images collected under severe weather such as fog and haze will suffer from quality degradation due to atmospheric scattering, which will make the image color off-white, reduce the contrast, and make it difficult to identify object features. It can also lead to biased understanding of image content. Image defogging refers to the use of specific methods and means to reduce or even eliminate the adverse effects of suspended particles in the air on images. Single image defogging refers to dehazing to obtain a clear image under the condition of only one foggy image. [0003] At present, single image defogging methods are mainly divided into three categories: the first category is based on image enhancement methods,...

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): G06T5/00G06T5/40G06T7/90G06N3/08G06N3/04
CPCG06T5/40G06T5/002G06T7/90G06N3/08G06N3/045
Inventor 苏延召崔智高李爱华王涛姜柯蔡艳平冯国彦李庆辉
Owner ROCKET FORCE UNIV OF ENG