基于龙芯和RTEMS的嵌入式图像处理平台的设计与实现

发布时间：2018-09-06 20:10

【摘要】：近年来,嵌入式系统发展迅速,应用广泛,在科技和生活中扮演了越来越重要的角色。相应的,嵌入式图像处理也在航空航天、智能机械等领域有了长足的发展。随着微处理器技术的迅速发展,带有操作系统的开发板越来越多,这种开发板既有较快的速度,也具有较高的通用性。然而,基于此的图像处理平台方面的研究仍处于比较欠缺的阶段。针对以上问题,本文结合了图像处理与嵌入式系统开发的相关技术,利用较成熟的嵌入式开发板及嵌入式操作系统为基础,根据其特性,设计实现了一种基于嵌入式操作系统的图像处理平台。该平台旨在搭建一个可用的嵌入式平台,使之能完成一系列图像处理功能,并将这些功能以接口的形式提供给后续的开发人员。基于此需求,本文首先进行了 RTEMS在龙芯1B上的移植工作,以支持上层软件接口的进一步开发,移植工作主要体现于板级支持包的实现。接着,本文进行了平台的功能接口的实现,本平台提供的图像处理功能模块接口主要包括单帧图片获取、图片处理、图片存取和图片传输这四大部分,其中,图片处理是指对图片进行灰度化处理、缩放、去噪等操作,在此功能中,针对于嵌入式平台的特性对一些图片处理方案做了适应性的改进应用,使之更好的适应于嵌入式平台。最后,选取了较具有代表性的交通标志识别场景对此平台进行了应用和测试,基于场景进行平台的功能接口的选取,并加以利用,提供了一个可用的系统例程。该平台使嵌入式图像应用开发的过程趋于简易,缩短了开发周期,并提高了用户友好性。课题完成了基于RTEMS和龙芯的嵌入式图像处理平台的开发,丰富了嵌入式图像处理的实现途径,提供了更有效的嵌入式图像处理方式。
[Abstract]:In recent years, embedded system is developing rapidly and widely used, which plays a more and more important role in science and technology and life. Accordingly, embedded image processing has made great progress in aerospace, intelligent machinery and other fields. With the rapid development of microprocessor technology, there are more and more development boards with operating system. However, the research of image processing platform based on this is still in a relatively short stage. In view of the above problems, this paper combines the related technologies of image processing and embedded system development, using more mature embedded development board and embedded operating system as the basis, according to its characteristics, An image processing platform based on embedded operating system is designed and implemented. The platform aims to build a usable embedded platform that can complete a series of image processing functions and provide these functions to subsequent developers in the form of interfaces. In order to support the further development of the upper software interface, this paper firstly carries out the transplantation of RTEMS on Longson 1B, which is mainly reflected in the implementation of board support package. Then, this paper implements the functional interface of the platform. The image processing module interface provided by the platform mainly includes four parts: single frame picture acquisition, picture processing, picture access and picture transmission. Image processing refers to the grayscale, scaling, denoising and other operations of the picture. In this function, some image processing schemes are adapted to the embedded platform in order to adapt them to the embedded platform. Finally, a representative scene of traffic sign recognition is selected to test and apply the platform. Based on the scene, the functional interface of the platform is selected and utilized, and a system routine is provided. The platform makes the process of embedded image application development easy, shortens the development cycle, and improves user friendliness. The project has completed the development of embedded image processing platform based on RTEMS and Godson, which enriches the implementation of embedded image processing and provides a more effective embedded image processing method.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP391.41;TP368.1

【参考文献】