摘要
本文针对复杂背景下目标的自动提取与识别这一难题,提出了一种基于知识的双特征融合多窗口结构识别算法。该算法包括粗检和细检两大步骤,即整幅图像的初步识别和感兴趣窗口内的双特征融合识别。依据一定的先验知识和准则,辅助以人工智能的手段,对复杂背景下的目标进行提取、复合、识别一条龙处理,从而达到识别出较为完整的目标的最终目的。
A knowledge-based bino-feature-fused multi-window architecture algorithm for targets extraction and recognition under complex background is proposed. The algorithm consists of two steps: preliminary recognition for the whole image and bino-feature-fased recognition for the interesting windows. With the help of certain prior knowledge, criteria and means of artificial intelligence, targets are extracted, fused and recognized quite well.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
1995年第3期237-242,共6页
Pattern Recognition and Artificial Intelligence
关键词
先验知识
目标识别
图像识别
人工智能
Prior Knowledge, Multi-window Architecture, Bino-feature-fused Algorithm, Perception Organization.
