生物地理学优化(BBO)在图像分割中的应用

Application of Biogeography-Based Optimigation to Image Segmentation

引用生物地理学优化的思想解决了图像分割的问题,该方法简称BBOIS。它采用图像像元的分形维、能量和灰度特征进行图像的初始分割,将一幅图像划分为3幅二值图像,作为初始分割图像("住地"),经像元的迁出、迁入的交换实现初始分割图像的优化。在优化过程中,还引用反向学习技术(OBL)提高分割图像的质量。3幅航空影像(100像元×100像元)的实验表明,BBOIS算法有一定的优越性,是有一定潜力的分割方法。

This paper puts forward a new application for image segmentation with the concept of biogeography-based optimigation（BBO）.It divides one original image into three binary images to be the initial segmented images by the fractal dimension,energy and gray of each pix in original image（residence）,and then reaches the image segmentation optimigation by the emigration and immigration of the pixels in the three initial segmented images.With the application of opposition-based learning（OBL） technology during the optimigation can also better the performance of image segmentation.The result of image segmentation experiment based on three aerial images（100 pixel×100 pixel） shows that the BBOIS algorithm can be proved to be advantage and potential.