【作者】 薛卫鹏；

【导师】 高立群；

【作者基本信息】 东北大学 ， 导航、制导与控制， 2009， 硕士

【摘要】 图像分割是计算机视觉领域的一个重要而基本的问题；是图像处理、分析和理解等领域中的一项关键技术。图像分割的研究一直都是图像处理技术研究中的热点和焦点之一。图像分割的算法层出不穷,在计算机视觉、模式识别和医学图像处理等实际中得到了广泛的应用。阈值图像分割方法是最常用的图像分割方法。它通过选取阈值将图像分为目标和背景两部分,其关键技术是阈值的选取。最大类间方差(Otsu)法是常用的阈值图像分割方法之一,基于直方图,通过考虑图像的灰度信息选取阈值。但是在图像信噪比低,灰度差异不明显,目标面积小等情况下,该算法效率会随之下降,甚至会产生错误的分割。针对Otsu分割算法的不足,根据不同军事目标红外图像的不同特点采用了不同的分割方法与之结合,提出了改进Otsu算法。新算法不仅考虑图像的灰度信息,而且还考虑领域空间的相关信息的特点来保证图像分割精度,然后针对分割结果的空洞利用了数学形态学进行了处理,以达到最好的效果。文中对改进算法的有效性进行了实验,实验结果表明,新算法可行、可靠,达到了理想的分割效果,实现了比传统方法更快速、稳定的图像分割。更多还原

【Abstract】 Image segmentation is one of the basic and important techniques in computer vision, and it is a pivotal technique in image process, analysis and image engineering. The study on the image segmentation is always one of the key points in the image segmentation technique study, many image segmentation algorithms are advanced, and it is widely used in computer vision pattern recognition and medicine image processing.Threshold image segmentation method is in common use in the image segmentation. It distingusihes the image into objects and background by picking threshold, so the threshold picking is the key technique. The Otsu method is often used as one of the threshold image segmentation methods, which is based on histogram, chooses threshold by analysing image gray level information. But this method is ineffective when facing complex image, such as image with low SNR or gray level discrepancy and tiny area, the algorithms effect decrease, even occur errors.Aiming at the shortage of Otsu segmentation algorithm, this paper studies different algorithms, focuses on different military objects’infrared image segmentation methods that combine with the Otsu method, and advances a improved method. The new method not only considers the gray information of the image, but the related information of domain space to ensure the accuracy of image segmentation. And then aiming at the hole in the segmentation result, this paper fills the hole with Mathematic Morphology to get the best results.This paper does the experiment with the improved method for the validity, as the result shows, this algorithm is feasible and credible,and gives an ideal image segmentation outcome, and it is more fast and steady than conventional method.更多还原