基于火焰图像显著区域特征学习与分类器融合的回转窑烧结工况识别
发布时间：2018-07-23 13:07
【摘要】：提出一种基于火焰图像分析回转窑烧结工况的新方法.针对现有方法难以提取具有区分性和鲁棒性局部形态特征的问题,并考虑到各显著区域图像数据具有不同的统计特性,首先采用卷积独立子空间分析网络的方法逐层学习复杂性递增的各显著区域局部形态特征;然后采用单词包模型学习各显著区域全局形态特征;最后,采用基于随机向量函数连接网络和模糊积分的方法设计集成分类器.实验结果表明了所提出方法可以提高工况的识别精度.
[Abstract]:A new method for analyzing sintering condition of rotary kiln based on flame image is proposed. To solve the problem that the existing methods are difficult to extract local morphological features with distinction and robustness, and considering the different statistical characteristics of image data in each significant region, Firstly, the convolutional independent subspace analysis network is used to study the local morphological features of each significant region with increasing complexity, and then the global morphological features of each significant region are studied by using the word packet model. An integrated classifier is designed based on random vector function linking network and fuzzy integral. The experimental results show that the proposed method can improve the identification accuracy.
【作者单位】： 东北大学流程工业综合自动化国家重点实验室;
【基金】：国家自然科学基金项目(61273177,61573090,61573364)
【分类号】：TP391.41
[Abstract]:A new method for analyzing sintering condition of rotary kiln based on flame image is proposed. To solve the problem that the existing methods are difficult to extract local morphological features with distinction and robustness, and considering the different statistical characteristics of image data in each significant region, Firstly, the convolutional independent subspace analysis network is used to study the local morphological features of each significant region with increasing complexity, and then the global morphological features of each significant region are studied by using the word packet model. An integrated classifier is designed based on random vector function linking network and fuzzy integral. The experimental results show that the proposed method can improve the identification accuracy.
【作者单位】： 东北大学流程工业综合自动化国家重点实验室;
【基金】：国家自然科学基金项目(61273177,61573090,61573364)
【分类号】：TP391.41
【相似文献】
相关期刊论文 前10条
1 李树涛,唐艳,王耀南;基于色彩通道融合的火焰图像分割[J];湖南大学学报(自然科学版);2001年04期
2 郑道昌,王颖,冯志敏,胡海刚;火焰图像的数字化探讨[J];中国航海;2001年02期
3 王亚军;徐大芳;陈向成;杨圣;;基于火焰图像动态特征的火灾识别算法[J];测控技术;2007年05期
4 刘明君;;基于火焰图像的工业炉氮氧化物排放检测技术[J];软件导刊;2012年03期
5 谭海涛;刘立;;锅炉火焰图像特征分析[J];智能计算机与应用;2013年01期
6 龚志红;谭海涛;;基于炉口火焰图像的炼钢终点研究与优化[J];电脑知识与技术;2013年16期
7 朱树先;朱学莉;;一种改进的火焰图像分割方法[J];科技资讯;2011年34期
8 何建东;汪亚明;郑俊褒;;基于自由火焰图像的三维温度场重建方法研究[J];工业控制计算机;2013年03期
9 张红亮;李R，
本文编号：2139507
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2139507.html
