基于图像显著性检测的图像分割

图像分割在许多图像处理和机器视觉问题中是一个非常重要的过程,是将一幅图分割成几个显著的区域,然而不能将其中最显著的目标直接分割出来,需要进一步处理。为此本文采用显著性检测的算法实现了对目标的分割。显著性区域检测可以应用于目标检测、图像检索、图像分割等机器视觉问题。使用杨等人提出的基于图论的流形排序算法检测显著性算法得到显著性图,再结合mean-shift分割算法,实现了对视觉显著性目标分割提取,可获得可观的图像分割结果,并将此算法应用到了森林火灾检测中,能对图像中的火焰部分进行有效的分割提取。

Image segmentation based on saliency detection

Image segmentation, the process of breaking a given image into salient regions, is an important process in many image processing and computer vision problems. However, it cannot get the most salient region, which should be handled further. A saliency detection algorithm is therefore applied to support image segmentation. Saliency detection can be applied to many computer vision problems, such as object detection, image retrieval, and image segmentation. We segment the salient object by the saliency detection algorithm via graph based manifold ranking algorithm proposed by Yang et al combined with the mean shift segmentation algorithm. Experimental results show that the results are impressive and this algorithm can be applied to forest fire detection, in which the fire part in the image can be segmented effectively.