基于重要度理论的图像识别方法  

Method of image recognition based on importance theory

作　　者：郭松 范存群[2] GUO Song;FAN Cunqun(College of Communication, Hulunbuir University, Hulunbuir 021008, China;National Satellite Meteorological Center,China Meteorological Administration,Beijing 100081, China)

机构地区：[1]呼伦贝尔学院传媒学院,内蒙古呼伦贝尔021008 [2]中国气象局国家卫星气象中心,北京100081

出　　处：《计算机集成制造系统》2021年第9期2736-2740,共5页Computer Integrated Manufacturing Systems

基　　金：国家重点研发计划资助项目(2018YFB0505000)。

摘　　要：为解决图像识别准确性问题并提升识别效率,提出一种基于重要度理论的图像识别方法。首先给出重要度理论的相关定义和定理,然后根据图像特征的映射建立图像特征知识域,并在此基础上计算得到图像特征知识的重要度,最后将图像特征知识的重要度作为权值进行图像特征向量的加权模计算,从而对图像进行识别。通过实验仿真证明所提方法可有效地识别图像,且具有较高的计算效率。Importance is a newer method of computation in rough set theory and reflects the degree to which an attribute affects a domain.An image recognition method based on importance theory was proposed to improve the accuracy of image recognition and ensure the timeliness of algorithm calculation.The definition and theorem of importance theory was given,and then the image feature knowledge domain was built according to the mapping of image features.On this basis,the importance of image feature knowledge was calculated.The importance of image feature knowledge was taken as the weight of image feature vector calculation to identify the image.Simulation results showed that the proposed method could effectively identify the image,and the algorithm had high computational efficiency.

关 键 词：图像识别 重要度 知识域 特征向量 权值 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]