一种改进的模板匹配方法在弹点识别中的应用

Application of An Improved Template Matching Method in Bullet-spot Recognition

在自动报靶系统中,使用模板匹配方法找出真实弹点是报靶的关键所在。但 传统的模板匹配方法受对象物大小变化、旋转、扭斜等变动影响,需要准备各种可能 的变形弹点模板,而且运算量大,难以提高识别效率。本文根据模板匹配方法的基本 原理,结合弹点的统计性质特征,采用改进的模板匹配方法来识别弹点,利用弹点特 征向量固定的这个前提条件,极大地减少了弹点识别需要的耗时,提高了自动报靶系 统的工作效率。

In an automatic target-scoring system, finding real bullet-spots by a template matching method is critical for target-scoring. However, traditional template matching methods would be affected by the size variation, rotation and deflection of objects. So, various possible targetscoring templates should beprepared. This would not only lead to a large quantity of operation, but also make it difficult to improve recognition efficiency. According to the basic principle of the template matching method, an improved template matching method for recognizing bullet-spots is put forward in this paper. Because the statistical property of bullet-spot recognition is incorporated and a precondition of the fixed bullet-spot eigenvector is used, it can greatly decrease the time for bullet-spot recognition and improve the efficiency of automatic target-scoring systems.