摘要
在工程中 ,实现基于人工神经网络的MRTD客观测量方法 (即以计算机对红外图像的自动智能客观判断来代替人眼的主观判断 )中 ,确定四杆靶图像的条纹对比度和均值对比度作为训练和识别的特征向量。讨论了特征向量选择的合理性 ,并给出了易于计算机实现的算法 。
In the engineering realization of objective measurement of MRTD which is based on artificial neural network (viz.subjective judgment of eyes is replaced by intelligent objective judgment of computer in the detecting of infrared image),contrast of tripe and contrast of mean are educed as the eigenvector of training and identification.The rationality of eigenvector as the choice is discussed and the algorithm that can be realized easily by computer is offered.The experiment result illuminates that the defining of eigenvector is reasonable and feasible.
出处
《激光与红外》
CAS
CSCD
北大核心
2002年第2期104-106,共3页
Laser & Infrared