一种融合区域生长和边缘检测的彩色图像分割方法

介绍了一种融合边缘检测和区域生长的彩色图像分割方法。算法首先对图像进行边缘检测,然后计算像素之间在HSI彩色空间的颜色相似度,利用像素间相似度并结合边缘信息确定区域生长的种子,再基于颜色和空间信息对各个种子进行区域生长,最后进行区域合并。该算法能够实现种子的自动选取,这在传统方法中是很难实现的,此外,还能有效防止过分割。实验结果表明该算法有效。

Color Image Segmentation by Combining Region Growing and Edge Detection

In this paper, a color image segmentation method which combines edge detection and region growing is proposed. The edge of color image is detected first, and then seeds are automatically selected depending on calculated pixel intensity similarity of pixels in the HSI color space combined with the edge information. Seeded region grows based on color and spatial information. The presented method can select the seeds automatically, as contrasts with traditional ones. Moreover, it can avoid over segmentation. The experimental results show that the method is effective.