海量弱关联图像数据的检测系统设计与实现  

Design and implementation of massive weak-correlation image data detection system

作　　者：李冰[1] 毛布 许然[3] 

机构地区：[1]四川商务职业学院信息技术系,四川成都610091 [2]自贡广播电视大学,四川自贡643000 [3]浙江理工大学,浙江杭州310018

出　　处：《现代电子技术》2017年第8期114-116,共3页Modern Electronics Technique

基　　金：国家自然科学基金(50875245)

摘　　要：当前设计的图像数据检测系统大都依据特定关键词完成数据检测,对海量弱关联图像数据的检测效率和精度较低。为了解决这些问题,设计基于PMML规范的海量弱关联图像数据检测系统,其将数据检测算法同PMML标准融合,由表示层、应用服务层和数据资源层构成。其中应用服务层中的数据检测模块将形成的检测模型通过PMML文档输出,模型应用模块通过PMML模型采用自定义的数据检测查询语言CMQL完成语言检索,实现海量弱关联图像数据的有效检测。系统实现部分给出了PMML模型的创建和应用过程。实验结果表明,所设计检测系统具有较高的检测性能和较低的虚警率。The currently-designed image data detection systems detect the data mostly according to the specific keywords, and have low detection efficiency and precision for the massive weak-correlation image data. In order to solve these problems, a massive weak-correlation image data detection system based on PMML standard was designed, which fused the data detection algorithm with the PMML standard. The system is composed of the presentation layer, application service layer and data resource layer. The data detection module of the application service layer outputs the formed detection model through the PMML document. The model application module uses the self-defined data to detect the query language CMQL through PMML model for language retrieval, so as to detect the massive weak-correlation image data effectively. The creation and application process of the PMML model is given in the fourth part of this paper. The experimental results indicate that the detection system has high detec- tion performance and low false alarm rate.

关 键 词：海量弱关联图像数据 PMML CMQL 数据检测系统 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置] TN453[自动化与计算机技术—控制科学与工程]