摘要“图代炮”系统，即对自行反坦克炮瞄准镜视场图像分析与评价的系统，从而代替了实弹射击训练。由跟踪瞄准图像摄取分系统、数据采集与保存系统和图像分析处理与评价系统三部分构成。图像分析处理与评价系统是本文研究开发的核心模块。对从瞄准镜视场摄取的全黑（或全白）的标准靶图像分析处理，判读瞄准标志的大立标箭头与目标相对位置，判定发射火炮是否命中。

图像处理包括大立标箭头的识别和目标靶的识别。在大立标箭头识别中，应用Hough变换及图像几何变换预处理图像，简要分析了各种模板匹配算法的优缺点，并提出了一种对二值图像进行模板匹配的算法，在匹配速度上具有良好效果。在目标识别方面，先对目标进行边缘检测，并进行一系列的图像处理后，找出靶面外加矩形，进行大小检测。根据靶面和大立标顶点的相对位置可以判断出命中与否。67757

图像处理是在VC++编译环境下基于OpenCV类库库实现系统的基本功能，本课题的研究及众多实验图像的综合处理，表明：本软件能够对2000米射程内的大立标箭头和目标靶进行比较准确的定位，同时满足实时性的要求，在特定情况下，基本实现了自动检测、判读的功能。

毕业论文关键词：图像分割，特征识别，模板匹配，图像判断，OpenCV类库

毕业设计说明书（论文）外文摘要

Title    Image recognition system design based on feature extraction                     

Abstract

"Figure generation of gun system, namely the at gun sights on its own field of image analysis and evaluation of the system, instead of the live firing training. By tracking target image intake subsystem, data acquisition and preservation system and the analysis of image processing and evaluation system of three parts. Image analysis processing and evaluation system is the core of this paper the research and development module. From the sight view to absorb all black (or white) standard target image analysis processing, interpretation aiming symbol of great beacon arrow as opposed to a target position, determine whether the fired artillery hit.

Image processing including large beacon arrow target identification and target recognition. In large beacon arrow recognition, and image Hough transform is applied to preprocessing the image geometric transformation, briefly analyzes the advantages and disadvantages of all kinds of template matching algorithm, and proposes a kind of binary image template matching algorithm, has a good effect on the matching speed. In terms of target recognition, first to edge detection of target, and make a series of image processing, to find a target surface with rectangle, detection of size. According to the relative position of target surface and the big beacon vertex can determine hit or not.

Image processing is under the environment of vc + + compiled based on the OpenCV library library to realize the basic function of system, this topic research and comprehensive treatment of numerous experimental image, shows that this software can be the great beacon arrow of the 2000 m range and target to compare accurate target positioning, at the same time meet the requirements of real-time, in certain cases, the basic realization of the automatic testing and interpretation.

Key words: Image segmentation, character recognition, template matching, image, the OpenCV library

目   录

1  绪论 1

1.1  课题的应用背景及意义 1

1.2  国内外研究现状