针对基于内容的图像检索与识别在智能监控中应用的问题,建立了基于区域划分的颜色链码,利用图像颜色的区域特征值来构成一个区域与整体的特征描述,并将它应用到捕获图像的识别。捕获的图像,通过梯度特征对来定位与检测,对获取的脸部图像进行链码特征匹配与识别,并将计算结果作为监控系统的判断与决策依据。运用这一方法的监控系统,提高了监控系统的辨别对象能力,拓展了智能监控研究领域。
This paper presents an application with content-based image retrieval in intelligent monitor and operator. Usually, color is used to describe the image, but not exquisite in detail of image. The chain coding is built by eigenvalue of region description that is partite from image. The grid feature couple is used to locate and detect the human face for the captured image. Extraction the feature and calculating the chain coding, the similarity function can decide the match degree for the processing object. This system enhances the intelligence in how to identify the object when the camera captures its image.