基于SRG的快速图像分割算法研究与实现  

Research and implementation of a rapid image segmentation algorithm based on SRG

作　　者：唐万成[1] 叶邦彦[1] 梁立东[1] 

机构地区：[1]华南理工大学机械与汽车工程学院,广东广州510640

出　　处：《电子设计工程》2014年第22期177-180,共4页Electronic Design Engineering

基　　金：国家自然科学基金资助项目(50875089)

摘　　要：针对传统的SRG算法实时性和准确性不足,提出了一种模拟自然界生态中病毒感染过程的智能快速图像分割算法,并在VC++6平台上实现。该算法基于传统SRG图像分割原理,一方面可以自动探测生长点,并且可根据背景的亮暗动态更新参数,对光线不均匀的图像也可进行分割;另一方面仅通过一次对图像的全局遍历搜索,便可逐一将目标从背景中分割并且作标记。实验结果表明,与传统阈值算法相比,该算法在提高图像分割速度的同时,在不同背景中分割图像的准确性更高。In order to improve the real-time and accuracy of the traditional SRG algorithm in image processing, this paper proposes a smart and rapid image segmentation by imitating the process of the virus infection in nature, and then implement it on VC＋＋6 platform. On one hand , The Algorithm can automatically detect the seeds in image region and can be adapt for uneven-light image by adjusting the parameters based on the brightness of the background; On the other hand, only by one of the image scanning, it can segment and label the objects from the background. The experimental results show that, compared with the traditional SRG Algorithm, this algorithm improves the segmentation speed in different background with higher accuracy.

关 键 词：SRG 图像分割 VC++6 光线不均匀 智能 

分 类 号：TN919.8[电子电信—通信与信息系统]