首页> 中文期刊> 《红外技术》 >融合直方图高阶统计特征与GLCM特征的室内红外图像人群密度分类

融合直方图高阶统计特征与GLCM特征的室内红外图像人群密度分类

         

摘要

公共场所的人群密度信息在公共安全、交通管理、应急减灾等方面具有重要作用,采用红外技术,可以在拍摄人群图像时避免环境光照影响.为了实现室内场景下的红外图像人群密度分类,提出一种融合灰度直方图高阶统计特征与灰度共生矩阵特征的人群密度分类方法.首先,根据红外图像的特点,分析并提取样本图像灰度直方图的高阶统计特征,随后与提取的灰度共生矩阵特征串行融合,最后作为多分类支持向量机的输入,对不同人群密度等级进行分类.实验结果表明,提出的方法对于不同密度人群图像的分类准确率可达92.13%,同时特征向量提取简洁、算法耗时短.%The crowd density information in public places plays an important role in public safety, traffic management, and disaster reduction in emergencies. The use of infrared technology can avoid the influence of ambient light while capturing crowd images. In order to realize indoor crowd density classification in infrared images, this paper proposes a method that fuses high-order statistics of a grayscale histogram with gray level co-occurrence matrix features (GLCM). First, considering the characteristics of infrared images, this paper analyzes and extracts the high-order statistics of the grayscale sample image histograms. Next, the histogram and GLCM features of sample images are fused serially. Finally, the fusion feature is input to the multi-class support vector machine and the classified crowd density level is output. The experimental results show that the proposed method can achieve 92.13% accuracy for different crowd density classifications in infrared images, with fewer features in less time.

著录项

相似文献

  • 中文文献
  • 外文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号