基于MATLAB的靶纸图像识别研究
2016-12-19王蔚扬丁嘉月汪鹏洪卢正勇
王蔚扬 丁嘉月 汪鹏洪 卢正勇
摘 要: 针对射击比赛中需要实时并且准确地对射击结果进行检测输出,提出了一种基于MATLAB函数工具箱的靶环数识别方案。采用数码相机拍摄靶纸图像,利用基于MATLAB的图像处理算法对其进行处理从而得出弹着点环数。对靶纸图像进行灰度转换、中值滤波、二值化等预处理,并采用Sobel算子进行边缘提取,减少后续图像处理的数据量。通过靶纸圆环的对称特征,先提取出圆心,将Hough变换降维来求内圆环参数,再结合图像差影法以及形态学处理对弹孔识别定位。实验发现,上述图像处理算法有较快的速度和较好的准确度。
关键词: MATLAB; 预处理; Hough变换; 降维; 形态学
中图分类号:TP391.41 文献标志码:A 文章编号:1006-8228(2016)11-17-04
Research on the image recognition of target sheet with MATLAB
Wang Weiyang, Ding Jiayue, Wang Penghong, Lu Zhengyong
(School of Imformation & Electronic Engineering, Zhejiang Gongshang University, Hangzhou, Zhejiang 310018, China)
Abstract: According to the shooting competition when a real-time and accurate detection of the shooting results is required, a research on image recognition of target sheet with MATLAB is presented in this paper. An image processing algorithm based on MATLAB is used in order to analyse the ring number of impact point on the target sheet captured by digital camera. Grayscale conversion, median filtering, and binariztion of the target sheet image are used in the image pretreatment. Sobel operator is used to extract image edge, so the amount of computation is dramatically reduced. And the bullet hole is recognized and located by using improved Hough transform, image subtraction method and morphological algorithm. The experimental results show that the image processing algorithm has fast speed and good accuracy.
Key words: MATLAB; pretreatment; Hough transform; mension reduction; morphology
0 引言
随着科技的发展,射击比赛时采用自动报靶系统,将会越来越普遍。这种方式不仅能减少人力物力支出、提高报靶效率,还能减小由于人为带来的不确定因素。目前已有不少单位研究出多种自动报靶系统[1-4]。基于机器视觉的自动报靶系统,应用环境要求特殊,还未得到普及[1-4]。本文基于MATLAB函数工具箱实现靶纸弹孔检测的数字化方法,在射击时对靶纸图像进行采集,通过基于MATLAB的图像处理技术对靶纸图像进行处理分析,并模拟图像进行成绩输出。
1 靶纸图像采集读取
靶纸图像采集可以有多种方式,本实验采用最高像素值为300万的数码相机拍摄,图像采集简便,便于推广应用。MATLAB中提供的imread()函数用于实现图像的读取操作,可读取JPG、TIF、GIF、HDF、XWD和CUR等多种图像格式。……
