摘要
针对红外图像中行人目标分割时容易受环境干扰的问题,提出一种基于马尔可夫随机场的红外行人分割方法。为了解决计算最优解时优化算法收敛速度慢的问题,采用simple linear iterative clustering算法将红外图像预分割为超像素块。以像素块作为马尔可夫随机场的分割单位。像素块的颜色信息由该像素块内所有像素的平均灰度值表示。根据马尔可夫随机场与图像信息之间的关系计算图像的局部先验空间结构信息,采用ICM计算后验能量函数的最小化解。实验结果表明,改进分割算法较好的抑制了小的噪声点,获得了比较理想的分割结果,对于360×240的红外图像,采用超像素对图像进行预分割可以将分割速度提高10倍以上。
Because the segment of pedestrian in infrared image is vulnerable to the environment,this paper presented an infrared pedestrian image segment algorithm based on markov random field.For the purpose of advancing the convergence speed,this algorithm adopts the simple linear iterative clustering algorithm to pre-segment the image into several super-pixels.And these super-pixels are the process units of the markov random field.The color information of the super-pixel is the average gray level of the pixels in it.The local priori distribution of the image is described according to the relationship between markov random field and image,and ICM algorithm is employed to optimize the posteriori energy function to get the optimal segmentation.The experimental results show that the algorithm in this paper can restrain the tiny noise better,so that the segment is perfect relatively.And the pre-segment of image can increase the computation speed more than 10 times for infrared images of size 360×240.
出处
《计算机仿真》
CSCD
北大核心
2012年第10期26-29,404,共5页
Computer Simulation
基金
国家自然科学基金(61170185)
辽宁省科技攻关计划项目(2011217002)
