摘要： 针对军警靶场对实弹射击结果高精确度和强环境适应性的要求,提出一种基于视频图像分析的自动报靶方法。提取不同环境下的靶图特征并使用Boosted Cascade算法训练靶图分类器,以提高报靶系统对环境的适应能力。对识别出的靶图做二值化与阈值分割处理,利用形态 学方法、霍夫变换、弹孔差影法求取靶心和弹孔位置,依据靶心坐标、环线半径以及弹孔坐标计算射击成绩。实验结果表明,该方法能正确识别靶图并计算射击成绩,对900次野外实弹射击的报靶精确度可达到98.5%以上。

关键词: 自动报靶, Boosted Cascade算法, 分类器, 霍夫变换, 弹孔识别, 军警靶场

Abstract: In order to satisfy the high precision and strong adaptability of the military riflery,this paper proposes an automatic target-scoring method based on video image analysis.It extracts the characteristics of the target figure under different environments to improve the system’s adaptability to the environment,and trains the classifier based on Boosted Cascade algorithm.It calculates the position of bull’s-eye and bullet holes through the binarization and threshold segmentation,morphology processing,Hough transform and bullet difference method.It applies the coordinates and radius to calculate the bullet shooting scores.Experimental results show that this method can quickly identify target figure and calculate the firing score correctly,and the statistical result of 900 wild live firing shoots shows that the precision can reach 98.5%.

Key words: automatic target-scoring, Boosted Cascade algorithm, classifier, Hough transform, bullet identification, military riflery

中图分类号: 

• TP391