首页|面向工业过程的图像生成及其应用研究综述

面向工业过程的图像生成及其应用研究综述

扫码查看
在面向工业过程的计算机视觉研究中,智能感知模型能否实际应用取决于其对复杂工业环境的适应能力.由于可利用的工业图像数据集存在分布不均、多样性不足和干扰严重等问题,如何生成符合多工况分布的期望训练集是提高感知模型性能的关键.为解决上述问题,以城市固废焚烧(Municipal solid wastes incineration,MSWI)过程为背景,综述目前面向工业过程的图像生成及其应用研究,为进行面向工业图像的感知建模提供支撑.首先,梳理面向工业过程的图像生成定义和流程以及其应用需求;随后,分析在工业领域中具有潜在应用价值的图像生成算法;接着,从工业过程图像生成、生成图像评估和应用等视角进行现状综述;然后,对下一步研究方向进行讨论与分析;最后,对全文进行总结并指出未来挑战.
Image Generation and Its Application Research for Industrial Process:A Survey
In computer vision research for industrial process,the practical implementation of intelligent perception models is contingent upon their capacity for adapting to complex environments.As a result of issues such as non-uniform distribution,inadequate diversity,and significant interference within available image datasets,generating a training set that meets the multi-condition distribution is pivotal to enhance model performance.In order to ad-dress these issues,with the municipal solid wastes incineration(MSWI)process as background,this article focuses on current research on image generation and its application for industrial process,providing support for perceptual modeling for industrial images.Firstly,the definition and process of image generation for industrial process are sum-marized,as well as their application requirements in industrial process.Subsequently,the image generation al-gorithms with potential application value in the industrial domain are analyzed.Then,an overview is provided from the perspectives of industrial process image generation,generated image evaluation and application.Next,the fu-ture research direction is discussed and analyzed.Finally,we summarize the article and provide future challenges.

Industrial processvisual perceptionimage generationgenerated image evaluation and applicationmunicipal solid wastes incineration(MSWI)

汤健、郭海涛、夏恒、王鼎、乔俊飞

展开 >

北京工业大学信息学部 北京 100124

北京工业大学智慧环保北京实验室 北京 100124

北京工业大学智能感知与自主控制教育部工程研究中心 北京 100124

工业过程 视觉感知 图像生成 生成图像评估与应用 城市固废焚烧

科技创新2030——"新一代人工智能"重大项目

2021ZD0112302

2024

自动化学报
中国自动化学会 中国科学院自动化研究所

自动化学报

CSTPCD北大核心
影响因子:1.762
ISSN:0254-4156
年,卷(期):2024.50(2)
  • 191