The interface recognition of inside computer is used in automatic test of radar countermeasure reconnaissance system. It can get high recognition rate using BP neural networks. First, the theory of BP neural networks is introduced, and then the paper gives the design method as a whole, and designs the server and the client. After that, the realization of recognition module is given. At last, the system is improved based on the trait of interface of inside computer, which improved the recognition rate. The entire system is efficient in speed, recognition and precision, the number recognition rate can reach 98 percent, and the letter recognition rate can reach 96. 5 percent. It has high prospect.%在对雷达对抗侦察装备进行整体性能自动测试时,是通过自动识别装备内部计算机界面的方法来实现的；采用BP神经网络算法可以得到很好的识别率；首先介绍了BP神经网络的识别原理,其次给出了系统的总体设计思路,对系统服务器和客户端进行了设计,并给出了识别模块的实现；最后针对装备界面显示的特点,对识别系统进行了改进,有效提高了识别率；整个系统速度快、识别效率高、准确性高,经实际测试,识别系统对于数字的识别成功率达到了98％,而对英文字母的识别成功率达到了96.5％,为装备的自动测试打下了很好的基础,具有一定的应用前景.
展开▼
