摘要
有两种方法可以用于多传感器系统的特征深度目标识别。第一种方法是基于静态算法的,而另外一种是基于神经网络的。第一种方法最主要的缺点是在目标识别的处理过程中依赖大量的假设;而第二种方法则不能处理复杂多传感器系统中的识别问题。而文中的神经网络组算法使神经网络具有了鉴定能力并保持了多传感器系统的优点。这种算法能解决在大噪音下的目标识别问题并提高了网络的速度。模拟体现了这种算法的成效。
There are two kinds of methods to characteristic-degree target recognition of multi-sensor system, as fidlows: the first method is based on statistic algorithm, the second is based on neural network. The first method's primary fault is depending on many assumptions during the process of target recognition, the latter can't deal with the identified problem to the complex muhi- sensor system. This paper interviews an algorithm based on neural network group, which has both the identified ability of neural net- work and the advantages of multi-sensor system. This algorithm can both solve the problem of identified targets with great noise and improve the velocity of network. The simulation shows the effectiveness of the algorithm.
出处
《微计算机信息》
2009年第31期152-153,共2页
Control & Automation
关键词
神经网络组
识别
多传感器
数据融合
neural network group
recognition
multi-sensor
data fusion