基于蚁群行为仿真的影像分割

基金项目: 国家自然科学基金资助项目(40271094)

详细信息

Image Segmentation Based on Simulation of Ant Colony Behavior

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

Funds: 国家自然科学基金资助项目(40271094)

摘要: 提出以像元的灰度值、能量值和分形维作为信息素的蚁群行为仿真的图像分割原理和方法。由该方法对3幅航空影像进行图像分割试验,得到图像的粗分割。利用本文提出的精处理方法,得到图像分割的最终结果。采用标记点量测的方法,对分割图像作出质量评价,结果表明,该方法很有应用前景。
- 蚁群行为仿真 /
- 图像分割 /
- 质量评价
Abstract: This paper presents a method and principle which employs pixel’s gray value, pixel's energy value and pixel's fractal value as pheromone to segment image through simulation of ant colony behavior. The method is tested on three aerial images. After the first process, the rough segmentation of image is produced. And the finial segmentation is achieved by using the elaborate method to process the rough results. In the end, quality of image segmentation is estimated by the means of measuring the matched point on image. The results expresses that what we put forward is promising method.
- simulation of ant colony behavior /
- image segmentation /
- quality estimate