摘要
提出了一种基于BP神经网络和DSmT推理的序列图像目标识别算法。以修正的Hu不变矩为图像特征,利用数据融合的思想对来自目标的序列图像进行时间域融合处理。由BP神经网络对目标的初步识别结果构造基本置信指派函数,用DSmT组合规则进行决策级数据融合,完成了三维飞机图像目标的识别仿真。仿真结果表明,融合方法提高了三维飞机目标识别的准确性。
A sequential image recognition algorithm based on BP Neural Network（BPNN） and Dezert-Smarandache Theory（DSmT） was presented. Firstly, modified Hu invariant moments was used as the feature of the image, and BPNN was applied to identify the target. Secondly, basic belief assignment function was constructed through the output of the BPNN, and afterwards DSmT fusion rule was used to finish the decision data fusion. Lastly, recognition of 3D airplane images was completed. The simulation results show that this fusion method is effective_
出处
《计算机应用》
CSCD
北大核心
2006年第1期120-122,共3页
journal of Computer Applications
