手持式检耳镜的视频流捕捉与处理

Capture and Process Video Stream of Handheld Otoscope

目的:医生借助医学图像观察病症时,医学图像的好坏直接影响医生对患者的诊断结果。外加人类的内耳结构相对身体其他结构较小,对图像的要求更加严格。在现有设备中并没有对输出图像的处理这个功能,从而导致一些疾病的误诊。本文中我们旨在增加图像后处理,提升图像的质量,使得显示出来的视频流数据能够更加清晰。方法:利用Visual Studio2010作为开发平台,建立上位机主窗口。基于第三方函数库Direct Show对视频流进行实时捕捉,并利用Open CV中的直方图均衡算法对视频进行后处理。利用两个程序得到不同的结果 ,一个在传输到上位机之前使用直方图均衡算法。另一个不使用直方图均衡算法,只是单纯将得到的图像输出,显示在PC机上。结果:将使用不同程序得到的两种图像进行对比,其中使用直方图均衡算法的图片,图像效果明显优于未使用算法的图像,特别是在病症处,病症区域的轮廓更加清晰,并且色彩更加柔和。结论:经过对比,我们可以明显看出使用直方图均衡算法的图像效果好,而且利用Open CV和Direct Show进行图像处理算法的编程,方法简单快捷,处理效果好,具有良好的跨平台移植性。

Objective：When doctors observed symptoms with medical image, the quality of the image will impact doctor diagnosis directly. Also the structure of human＇ s inner ear is smaller than other parts of human, so it will have stricter requirement in the image. But current device don＇t have the function of process the exporting image, thus it will course some misdiagnose. In this article, we are aim to add image post-processing to improve the quality of the image, make the video stream more clearly. Methods： We use Visual Studio2010 as a research platform to make a main window of PC. And we capture real time video stream base on DirectShow, use Histogram Equalization Algorithm which included in OpenCV to process the image. Then we use two different program, one add Histogram Equalization Algorithm before the video stream is transformed to the PC. The other don＇t have Histogram Equalization Algorithm, just display the video which we capture from the camera. Results： Compare the two different image which use two different program. The one which use Histogram Equalization Algorithm is obvious better, especially in the location of the disease. Conclusions： Based on the compare, we know the picture which use Algorithm is better, and the program that use OpenCV and DirectShow for image processing algorithm is easier on the method and treatment effect is good, also it has favorable portability of cross-platform.