摘要
Otsu最大类间方差法是图像阈值分割中的经典算法之一,在图像处理邻域应用广泛。为了改善传统二维Otsu算法存在运算时间较长,不适用于铁路轨道实时监测的缺点。文中将遗传算法与Otsu算法相结合,运用自适应遗传算法的原理,通过选择、交叉、变异等操作,寻找图像灰度最佳分割阈值。实验结果表明,在兼顾了图像分割质量前提下,文中算法大幅缩短了运算时间,较好提取了铁轨轮廓。
Otsu threshold algori thm is one ol the classic image segmentation. It is widely used in image processing,but Otsu image segmentation also has some disadvantages,such as more time-consuming,is unsuitable for railway tracks ’ real-time processing. Combing the AGA with two-dimensional Otsu algorithm is used to solve the optimum threshold,then through operations ol choice,cross mutation,the optimum threshold value is obtained. The experiment results indicated that the algorithm can extract the tracks contour effectively in consideration of the balance of quality and reduce the computation time.
出处
《信息技术》
2017年第4期125-128,共4页
Information Technology
关键词
铁路轨道
OTSU
遗传算法
阈值分割
railway tracks
Otsu
genetic algorithm
threshold segmentat ion
