摘要
针对铁路沿线护栏的检测问题,提出了一种改进的均值漂移算法。首先,从阈值分割算法入手,得到铁路防护栏网格区域的颜色范围;然后,基于均值漂移算法对其进行图像平滑和分割;再针对产生的过分割图像,采用基于最小面积的合并停止准则来进行区域合并;最后进行网格提取。实例验证表明:改进的均值漂移算法的效果优于传统的均值漂移算法。
Aiming at the detection of the railway guardrail,an improved mean shift algorithm was proposed in this paper. Firstly,the color range of the railway guardrail grid area was obtained from the threshold segmentation algorithm. Secondly,railway guardrail image smoothing and segmentation were conducted based on the mean shift algorithm. Furthermore,aiming at the generated excessive segmentation image,region merging was adopted based on the minimum area merging and stopping criterion. Finally,grid extraction was conducted. The experimental examples indicate that the effectiveness of the improved mean shift algorithm is better than that of the traditional mean shift algorithm.
出处
《河南科技大学学报（自然科学版）》
CAS
北大核心
2015年第3期28-32,6,共5页
Journal of Henan University of Science And Technology:Natural Science
基金
国家自然科学基金项目(5136501)
江西省自然科学基金项目(20132BAB203020)
江西省教育厅科学技术研究项目(GJJ13430)
关键词
铁路护栏
网格提取
阈值分割
均值漂移
区域合并
railway guardrail
grid extraction
threshold segmentation
mean shift
region merging
