为了能够安全、可靠地判别引信安全状态，实现自动化判别，首先在分析某型引信惯性保险机构特征的基础上，明确了保险钢珠的状态识别是关键；接着利用最优投影及形态学滤波技术，根据某型惯性保险引信X光图像中所反映的零部件位置、灰度和形态等特征，建立保险钢珠状态的识别算法；然后应用C语言进行算法编程，应用C++Builder完成程序界面设计，实现识别过程的人机交互。解决了惯性保险引信安全状态自动化判别的难题。

In order to judge the fuze safe condition safely and reliably and to realize the automatic and precise judgment, it was confirmed that the key was the safety pellet state identification at first. Then the optimal projection and morphological technology was adopted to constitute identification arithmetic of safety pellet state on the basis of characters of part location, gray scale and conformation represented by some type inertia safety fuze X ray pictures. Then the arithmetic was programmed by C-language, and the man-machine interface was designed by C++ Builder. Thus the problem of inertia safety fuze safe condition automatic and precise judgment was solved.