摘要:随着科技的发展,图像识别作为计算机视觉的一个重要研究领域,受到人们越来越多的关注。其中,将视频流与图像识别相结合,实现了对特定目标的实时监控,这种技术在农业生产,医疗救助,交通监管等方面都得到了广泛的应用。本文主要采用颜色识别的方法,通过在高速视频流中采集和筛选目标像素值及范围,精确识别各类目标并进行标注、显示和输出,包括位置、颜色、距离测算等,同时通过串口进行数据输出,并通过研究,提出了一种具有广泛应用领域的图像识别技术及相应的软件工具,为现代农业及工业领域的机器人视觉识别、目标定位等提供有价值的参考。29503 毕业论文关键词:视频流;图像提取;颜色识别;距离测算
Image target recognition based on video technology
Abstract: With the development of science and technology, image recognition as an important research field of computer vision, by people more and more attention. Among them, the video stream and image recognition combined to achieve a specific goal of real-time monitoring, this technology in agricultural production, medical assistance, traffic supervision and other aspects have been widely used. In this paper, the method of color recognition is used to collect and filter the target pixel values and ranges in the high-speed video stream, and the objects are identified and marked, displayed and output accurately, including location, color, distance measurement and so on, at the same time, through a serial port for data output, and through the research, put forward a wide range of application areas of image recognition technology and the corresponding software tools for modern agriculture and industrial field of robot vision recognition, target positioning and so provide a valuable reference.
Key words: video stream; image extraction; color recognition; distance measurement
目 录
摘要 1
关键词 1
ABSTRACT 1
KEY WORDS 1
1绪论 1
1.1研究背景 1
1.2国内外研究状况 2
1.2.1国内研究现状 2
1.2.2国外研究现状 2
1.3开发环境与技术介绍 2
1.3.1硬件环境 2
1.3.2软件配置 2
1.3.3 JavaCV概述 2
2课题概述 2
2.1课题设计 2
2.2颜色识别 3
2.3位置识别 3
2.4图像信息的获取与输出 3
3颜色识别 4
3.1图像预处理 4
3.2颜色空间转换 5
3.3目标筛选及跟踪 6
3.4图像形态学处理与阈值分割 7
3.4.1图像形态学处理 7
3.4.2二值化 7
3.5绘制轮廓 8
3.6获取目标颜色分量值 8
4位置识别 8
4.1立体标定 9
4.1.1坐标系变换 9
4.1.2标定技术 10
4.1.3标定流程 12
4.1.4标定结果与误差分析 12
4.2立体校正 13
4.3目标测距 14
4.3.1单目标测距 14
4.3.2多目标测距 15
5图像信息的获取与输出 17
5.1信息获取 17
5.2信息输出 18
6实验结果 19
7总结与展望 22