中文摘要：

针对IRST组网多目标测向定位系统中存在的虚假点问题，提出分层关联的思想。首先对交叉点进行预处理，去除不符合空间逻辑的交叉点；其次进行角度门限粗关联，去除大部分虚假点，减少计算量；最终利用多维分配思想选出正确的关联组合。计算机仿真结果表明，该算法可较好地对共面多目标进行定位，并且计算量适中。

英文摘要：

With the problem that associating measurements will produce ghosts in IRST network cross location system, a new algorithm developed is to eliminate the ghosts in layers is concerned. First, pretreatment to remove illogical cross locations, then a part of ghosts will be eliminated by angle hypothesis test, and finally picking up true associations from the remaining candidate associations is formulated as a multiple dimensional assignment prob- lem. Simulation results show that the algorithm can locate multiple coplanar targets simultaneously, and the compu- tation burden is moderate.