低对比度目标自动识别技术研究  

Research on Auto Recognition Technology With Low Contrast Target

作　　者：孙晓明[1] 霍富荣[1] 田凡[1] 王文生[1] 

机构地区：[1]长春理工大学现代光学测试技术研究室,长春130022

出　　处：《光子学报》2007年第11期2153-2156,共4页Acta Photonica Sinica

基　　金：总装备部十五预研基金(41317030106)资助

摘　　要：提出了在实时光电混合联合变换相关器中应用拉普拉斯锐化技术.通过对联合变换功率谱面的锐化处理,提高了联合变换功率谱的杨氏条纹对比度,使得相关峰的对比度明显增强,从而提高了目标识别率.实验证明:利用该技术所能探测的目标对比度最低限度约为0 .5 %.作为实例,给出复杂背景下低对比度(0 .8 %)地面目标汽车的实验结果.The technology of the laplacian sharpening effect is applied in the real-time photoelectronic joint transform correlator. Through sharpening process of joint transform power spectrum plane, it improves Yongrs interference fringes contrast of joint transform power spectrum. The constrast of correlation peak is enhanced,and the ratio of target recognition is also improved greatly. The experimental results indicate that the lowest limitation of detected target contrast used by this technology is about 0. 5%. The experiment result of the low contrast（0.8%） ground target-automobile under the cluttered background is given.

关 键 词：联合变换相关器 拉普拉斯锐化 复杂背景 目标识别 

分 类 号：TP317[自动化与计算机技术—计算机软件与理论]