基于吸引子的图像分割新方法的研究

详细信息

Research About New Method of Image Segmentation Based on Attractor

¹School of Remote Sensing and Information Engineering,Wuhan University,129 Luoyu Road,Wuhan 430079,China

摘要: 提出了一个“边界点”的新概念,并引用正态分布中大于3倍标准差的事件为小概率事件,作为图像分割中边界点“捡起”或“卸下”的依据。与其他方法相比,能使“捡起”、“卸下”的决策计算大大简化。通过对两幅航空影像的分割实验表明,新方法的分割效果是好的,是一个很有实用潜力的图像分割方法。
- 图像分割 /
- 吸引子 /
- 边界点
Abstract: This paper puts forward a new concept about "boundary point".The event in normal distribution that greater than thrice standard deviation is used as a small probability event and takes them as the evidence of boundary point for "pick up" and "drop down" in image segmentation.Compared with other methods,it makes decision making computation of "pick up" and "drop down" simplified greatly.The experiments on two pictures of air image show that the segmentation effect of new method is effective and it is a method of image segmentation with great practical.
- image segmentation /
- attractor /
- boundary point