计算机工程 ›› 2018, Vol. 44 ›› Issue (12): 251-257.doi: 10.19678/j.issn.1000-3428.0049039

• 图形图像处理 • 上一篇    下一篇

基于Haar-PP混合特征提取的沉降颗粒识别研究

周冉,宋华珠,刘翔   

  1. 武汉理工大学 计算机科学与技术学院,武汉 430070
  • 收稿日期:2017-10-23 出版日期:2018-12-15 发布日期:2018-12-15
  • 作者简介:周冉(1992—),男,硕士研究生,主研方向为图像处理、数据挖掘;宋华珠(通信作者),副教授、博士;刘翔,硕士研究生。
  • 基金项目:

    国家科技基础性工作专项(2014FY110900)。

Research on Settlement Particle Recognition Based on Haar-PP Mixed Feature Extraction

ZHOU Ran,SONG Huazhu,LIU Xiang   

  1. School of Computer Science and Technology,Wuhan University of Technology,Wuhan 430070,China
  • Received:2017-10-23 Online:2018-12-15 Published:2018-12-15

摘要:

针对沉降颗粒识别与轨迹追踪中由于灰度模糊、变化造成的识别通过率低、准确度差的问题,对由沉降颗粒的个体形态、运动形态以及空间运动引起的动态灰度变化进行分析,完成激光散射像点图像灰度增强、像素特征转换、像素点连通性判断等处理过程,在此基础上提出一种基于Haar-PP混合特征的沉降颗粒识别算法。实验结果表明,与高斯混合模型相比,该算法有较高的识别率,且时间复杂度较低。

关键词: 沉降颗粒, 激光散射, 特征提取, 灰度增强, Haar-PP混合特征

Abstract:

To solve the problem that gray blurring and variation cause low recognition rate and poor accuracy for trajectory tracking and trajectory tracking,the processing of gray-scale enhancement,pixel feature conversion and pixel point connectivity of laser scattering image is completed by the analysis of the dynamic gradation changes caused by the individual shape,motion shape and spatial motion of the settled particles.A settlement particle recognition strategy based on Haar-PP mixed feature is proposed.Experimental results show that compared with the Gaussian mixed model,the algorithm has higher recognition rate and lower time complexity.

Key words: settlement particle, laser light scattering, feature extraction, grayscale enhancement, Haar-PP mixed feature

中图分类号: