基于颜色相似性度量的弹孔识别方法

针对现有自动报靶系统对弱小弹孔目标识别能力差的缺点提出了一种新的弹孔识别方法。该方法以彩色图像为数据源，在分析弹孔成像的基础上采用颜色相似性概念对弹孔的特征进行描述。然后，以3×3的方形区域为基本的比较单元对射击前后两帧靶面图像的变化情况进行量化表示，并提取出满足人眼识别阈值的区域。最后，依据对弹孔特点的分析结果从这些区域中提取出满足弹孔特征的区域。实验结果表明：和现有方法相比，该方法充分利用了彩色图像包含的丰富信息，对与背景颜色相似程度低于98%的弹孔实现了有效识别，且符合人眼视觉生理特性。

Bullet holes recognization based on color similarity measurement

In view of the shortage existing in the method for small bullet holes recognization,a new method which used color image as the data source was introduced in this paper.With analysis of image characteristics of bullet holes which were describled by color similarity, the template of 3×3 square area was bulit as the basic comparison unit to get the differences between the preshot image and the after-shot image, and then the areas where the threshold was satisfied the human eye recognition were extracted from which the bullet hole features were recognized by analyzing the image. Compared with the existing methods, the experimental esults show that the approach could utilize the information consisted in color image fully which not only can recognize the bullet holes which has 2% difference between background, but also meet the color visual characteristics.