摘要
针对路标图像容易受噪声干扰和路标的识别问题,提出了一种基于改进马尔科夫随机场的路标去噪识别算法,标识变量采用EM学习算法估计;算法分为两步:首先利用马尔科夫随机场对已经被噪声污染的图像进行建模恢复;其次,通过计算恢复后图像的不变矩对路标图像进行分类识别;仿真实验比较了中值滤波方法和高斯平滑两种方法,实验结果表明利用马尔科夫随机场的方法与其他两种方法相比,该新算法处理后的图像可以达到较好的降噪效果,提高恢复路标图像的清晰度,能够更好地识别出路标,体现出该算法的优越性和有效性。
Signs image susceptible to noise interference and to identify road signs problems, puts forward a signpost denoising based on Markov random field recognition algorithm, identify variables using EM algorithm to estimate study. Algorithm can be divided into two steps： first, use of Markov random field for image modeling recovery has been noise pollution; Secondly, through the calculation of moment invariants of recovered image classifying signposts image recognition. Simulation experiment compares the median filtering method and the gaussian smoothing, two methods, the experimental results show that the method of using Markov random field compared with other two methods, the processed image is the new algorithm can achieve a better noise reduction effect, improve the recovery road signs image clarity, better able to identify the road signs, reflecting the superiority and effectiveness of the algorithm.
出处
《计算机测量与控制》
北大核心
2014年第4期1199-1200,1204,共3页
Computer Measurement &Control
基金
新疆大学校院联合资助项目(XY110133)
国家自然科学基金项目(60865001)
