燃油喷雾图像分析及粒子库的生成与扩充
发布时间:2018-10-29 10:06
【摘要】:随着石油资源的日益枯竭和世界范围内空气污染的加剧,致使各国对汽车尾气排放的要求越加严格,也使车企更加关注汽车的经济性和排放性,而对燃油喷雾特性进行深入研究对发动机燃油经济性及排放性有着重要的意义。本文提出在对喷雾图像进行处理的过程中,将图像中的粒子提取出来,并建立和扩充相应的粒子库。在MATLAB/GUI平台上开发燃油喷雾图像粒子库管理系统,并将该管理系统与SQL Server粒子数据库进行连接。将发动机喷雾实验获取的喷雾图像导入喷雾图像粒子库管理系统中,使用区域生长算法对阈值处理后的图像进行粒子信息提取,并将提取的粒子添加到粒子库中,完成粒子库的生成。调用粒子库中的粒子对阈值处理后的图像进行匹配,使用区域生长算法对未匹配成功的新粒子进行信息提取,并将新的粒子添加到相应的粒子库中,完成粒子库的扩充。通过改进的SSDA算法与NCC算法的融合,得到一种新的匹配算法进行图像匹配。本文用Photoshop画出仿真粒子图像,并将图像导入喷雾图像粒子库管理系统中,对喷雾图像粒子库管理系统进行调试,其运行结果验证了喷雾图像粒子库管理系统的实用性。将喷嘴孔径为0.366 mm,喷油压力为20Mpa工况下的两张较清晰的喷雾图像(a)和喷雾图像(b)分别导入粒子库管理系统中,图像(a)生成了该工况的粒子库,图像(b)对该工况的粒子库进行了扩充,并统计出图像(a)和图像(b)中不同长短轴粒子的个数。
[Abstract]:With the increasing depletion of oil resources and the intensification of air pollution in the world, the requirements of automobile exhaust emissions become more stringent in various countries, and the automobile enterprises pay more attention to the economy and emissions of cars. The research on fuel spray characteristics is of great significance to engine fuel economy and emission. In this paper, in the process of spray image processing, the particles in the image are extracted, and the corresponding particle library is established and expanded. A fuel spray image particle library management system is developed on MATLAB/GUI platform, and the system is connected to the SQL Server particle database. The spray image obtained from engine spray experiment is imported into the particle library management system of spray image, and the particle information is extracted from the image after threshold processing by using the region growth algorithm, and the extracted particles are added to the particle library. Complete the generation of particle library. The particles in the particle library are called to match the image after threshold processing, and the region growth algorithm is used to extract the information of the new particle that has not been matched successfully, and the new particle is added to the corresponding particle library to complete the expansion of the particle library. Through the fusion of improved SSDA algorithm and NCC algorithm, a new matching algorithm for image matching is obtained. In this paper, Photoshop is used to draw the simulated particle image, and the image is imported into the spray image particle library management system, and the system is debugged. The running results verify the practicability of the spray image particle library management system. Two clear spray images (a) and (b) with a nozzle aperture of 0.366 mm, injection pressure of 20Mpa were imported into the particle library management system respectively. The image (a) generated the particle library of the working condition. Image (b) extends the particle library of this working condition and calculates the number of particles in image (a) and image (b) with different long and short axes.
【学位授予单位】:长安大学
【学位级别】:硕士
【学位授予年份】:2016
【分类号】:U464
本文编号:2297389
[Abstract]:With the increasing depletion of oil resources and the intensification of air pollution in the world, the requirements of automobile exhaust emissions become more stringent in various countries, and the automobile enterprises pay more attention to the economy and emissions of cars. The research on fuel spray characteristics is of great significance to engine fuel economy and emission. In this paper, in the process of spray image processing, the particles in the image are extracted, and the corresponding particle library is established and expanded. A fuel spray image particle library management system is developed on MATLAB/GUI platform, and the system is connected to the SQL Server particle database. The spray image obtained from engine spray experiment is imported into the particle library management system of spray image, and the particle information is extracted from the image after threshold processing by using the region growth algorithm, and the extracted particles are added to the particle library. Complete the generation of particle library. The particles in the particle library are called to match the image after threshold processing, and the region growth algorithm is used to extract the information of the new particle that has not been matched successfully, and the new particle is added to the corresponding particle library to complete the expansion of the particle library. Through the fusion of improved SSDA algorithm and NCC algorithm, a new matching algorithm for image matching is obtained. In this paper, Photoshop is used to draw the simulated particle image, and the image is imported into the spray image particle library management system, and the system is debugged. The running results verify the practicability of the spray image particle library management system. Two clear spray images (a) and (b) with a nozzle aperture of 0.366 mm, injection pressure of 20Mpa were imported into the particle library management system respectively. The image (a) generated the particle library of the working condition. Image (b) extends the particle library of this working condition and calculates the number of particles in image (a) and image (b) with different long and short axes.
【学位授予单位】:长安大学
【学位级别】:硕士
【学位授予年份】:2016
【分类号】:U464
【参考文献】
相关期刊论文 前10条
1 曹建明;;射流表面波理论的研究进展[J];新能源进展;2014年03期
2 刘颖;;计算机图像识别技术的发展现状与展望[J];考试周刊;2014年08期
3 窦华军;;基于数字图像处理技术分析[J];数字技术与应用;2013年12期
4 聂欢欢;伊磊;刘任平;;基于区域生长法提取二值图像中的连通区域[J];计算机时代;2012年06期
5 陈露晨;李光明;;快速最大熵多阈值图像分割算法[J];计算机应用与软件;2012年03期
6 汪丹平;唐勇;;基于直方图的迭代式自动阈值分割技术研究[J];软件导刊;2011年08期
7 李贤华;彭真;;基于JDBC-ODBC的数据库访问方法的设计与实现[J];计算机与现代化;2008年05期
8 孙卜郊;周东华;;基于NCC的快速匹配算法[J];传感器与微系统;2007年09期
9 张杰;张蕊蕊;胡卜元;白素芳;徐璇;;煤焦SEM图像的表面孔洞分形维数的Matlab实现[J];河北工程大学学报(自然科学版);2007年02期
10 曹建明;喷雾学研究的国际进展[J];长安大学学报(自然科学版);2005年01期
,本文编号:2297389
本文链接:https://www.wllwen.com/kejilunwen/qiche/2297389.html