摘要
为解决装甲车车载图像跟踪系统中对场景进行分类的问题,提出了一种根据模糊模式识别原理对场景图像进行分类的方法。首先通过建立场景图像的特征向量来对场景进行描述,并根据所选特征参数设计了矢量隶属函数,然后通过计算隶属函数的值来对场景图像进行分类。实验证明,场景分类的结果对跟踪系统中目标提取方法的选取,有较好的指导作用。
In order to classify the scene pictures in image tracking system on armored vehicle, this paper pats forward a classification method based on blur pattern recognition principle. It first constructs feature vector to describe the scene, and then designs vector membership function according to the selected feature parameters, finally classifies the scene pictures by computing the value of membership function. As experiment shows, the scene classification result provides a good guidance for selecting target extracting algorithm in image tracking system.
出处
《微计算机信息》
北大核心
2007年第34期220-221,259,共3页
Control & Automation
基金
国家自然科学基金项目(10376043A06)
关键词
场景分类
模糊模式识别
隶属函数
四叉树
scene classification, blur pattern recognition, membership function, quad tree