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

Aluminum electrolysis FireEye image inpainting method based on deep convolution generative adversarial network

A deep convolution, aluminum electrolysis technology, applied in the computer field, can solve problems such as instability of training GAN

Pending Publication Date: 2020-05-22
CENT SOUTH UNIV
View PDF15 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] However, training GAN needs to achieve Nash equilibrium. Sometimes it can be achieved by gradient descent method, but sometimes it cannot be achieved. Because a stable method to achieve Nash equilibrium has not been found, training GAN is unstable.

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
  • Aluminum electrolysis FireEye image inpainting method based on deep convolution generative adversarial network
  • Aluminum electrolysis FireEye image inpainting method based on deep convolution generative adversarial network
  • Aluminum electrolysis FireEye image inpainting method based on deep convolution generative adversarial network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0036] The invention is a repairing method for the blocked fire eye image.

[0037] 1. Simple preprocessing of Huoyan image

[0038] In order to solve the impact of different image shapes, only take the square Huoyan image with the center of the Huoyan as the center and a side length of 400p. Since the image is a binary image, you can directly obtain the lower left corner of the Huoyan image (xl, yl) and the coordinates (xr,yr) of the upper right corner. Then according to the formula (xc,yc)=((xl+xr) / 2,(yl+yr) / 2), the center coordinates of the fire eye can be obtained, and finally the coordinates of the lower left corner and upper right corner of the image are extracted as (xl-100,yl -100), (xl+100, yl+100).

[0039] Second, use the Wasserstein distance to define the loss of the generator.

[0040] The Wasserstein distance is also called the Earth-Mover (EM) distance, which is defined as follows:

[0041]

[0042] Π(P r ,P g ) is P r and P g The set of all possible ...

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 an aluminum electrolysis FireEye image inpainting method based on a Waserstein deep convolution generative adversarial network. In the actual aluminum electrolysis process, a large amount of carbon slag falls off in the electrolysis process, and flame interference exists, so that a large number of FireEye images cannot be recognized, and the accuracy of image recognition isgreatly reduced. The invention discloses an image inpainting method combining DCGAN and WGAN methods. The method comprises two parts: firstly, carrying out simple preprocessing of an image, taking asquare FireEye image with the FireEye as the center and the side length of 400 pixels, thereby preventing the interference features of other images; and secondly, for the model of the invention, the characteristics of DCGAN and WGAN are integrated, and a Wasserstein deep convolution generative adversarial network (W-DCGAN) model combined with the DCGAN and the WGAN is used. According to the method, a convolutional network feature extraction capability and a Wasserstein function are combined for training, a loss function is optimized by using an RMSProp optimization algorithm, and then a generator model part in the trained W-DCGAN is extracted for a new network structure for image inpainting.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to an aluminum electrolytic fire eye image repair method based on Wasserstein deep convolution generation confrontation network. Background technique [0002] The superheat of the electrolyte in an aluminum electrolytic cell refers to the difference between the electrolyte temperature and the primary crystal temperature. It is generally believed that it is more appropriate to maintain the superheat at about 8-12°C, which can not only maintain the normal progress of the aluminum electrolysis process, but also ensure low energy consumption. loss. Therefore, it is particularly important to judge the degree of superheat in the production process. At present, the degree of superheat is divided into three categories: high, normal and low. Maintaining a good degree of superheat can reduce the energy loss in the process of aluminum electrolysis. However, in the actual aluminum electroly...

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 Applications(China)
IPC IPC(8): G06T5/00G06N3/04G06N3/08
CPCG06N3/08G06T2207/20081G06T2207/20084G06N3/045G06T5/77
Inventor 陈晓方潘慕尧谢永芳谢世文
Owner CENT SOUTH UNIV