首页|基于多特征的地铁运行监控图像目标识别研究

基于多特征的地铁运行监控图像目标识别研究

扫码查看
地铁监控图像中的目标检测与识别是保障地铁安全运行的关键技术,但地铁监控环境复杂多变,会给目标识别带来挑战.为准确识别地铁监控图像中的目标,研究了一种基于多特征融合的方法.通过分析监控场景的特点,提出了融合颜色、纹理、形状等多种视觉特征的方案,设计了有效的特征融合策略,构建了基于深度学习的多特征目标识别系统.结果表明,不同特征的协同与互补不仅提高了识别准确率,也提高了对复杂环境的适应性.
Research on the Target Recognition of Subway Operation Monitoring Image Based on Multiple Features
Target detection and recognition in subway monitoring images are key technologies to ensure safe operation of the subway,but the complex and ever-changing subway monitoring environment can pose challenges to target recognition.To accurately identify targets in subway monitoring images,this paper studies a method based on multi feature fusion.By analyzing the characteristics of monitoring scenes,a scheme was proposed to integrate multiple visual features such as color,texture,and shape.An effective feature fusion strategy was designed,and a multi-feature target recognition system based on deep learning was constructed.The results indicate that the collaboration and complementarity of different features not only improve recognition accuracy,but also enhance adaptability to complex environments.

subway monitoringtarget identificationmulti-feature fusionmonitoring and security

杨焰鑫、张振业、张小宁、张博文、孙国康

展开 >

北京市轨道交通运营管理有限公司,北京 102628

苏州华兴致远电子科技有限公司,江苏苏州 215000

地铁监控 目标识别 多特征融合 监控安全

2024

自动化应用
重庆西南信息有限公司

自动化应用

影响因子:0.156
ISSN:1674-778X
年,卷(期):2024.65(4)
  • 2