基于图像分析的线缆节距测量算法研究  

Research on Cable Pitch Measurement Algorithm Based on Image Analysis

作　　者：石守东[1] 王刚[1] 

机构地区：[1]宁波大学信息科学与工程学院,浙江宁波315211

出　　处：《计算机工程》2015年第11期273-279,286,共8页Computer Engineering

基　　金：浙江省重中之重学科开放基金资助项目(XKXL1514);宁波大学王宽诚教育基金资助项目

摘　　要：随着线缆传输速度的不断提高,对线缆节距的测量精度提出了越来越高的要求。为此,设计一种基于图像检测技术的节距测量算法。分割线缆的前景和背景图像,根据前景像素点在特征空间中的分布情况,实现模糊C均值聚类。采用基于边缘定位的特定像素点填充、图像细化以及数据拟合技术,求得拟合方程的可行解以及相邻可行解之间的距离,通过图像比例尺将图像节距值转换为实际节距值。实验结果表明,该测量算法可快速有效地计算出线缆节距,且与激光测距法的绝对误差约为0.75%,具有较高的测量精度。With the improvement of cable transmission speed, higher and higher demands of measurement precision of cable pitch are proposed. For this reason, this paper designs a cable pitch measurement algorithm based on image detection technique. It splits foreground and background of the cable image and uses the distributed situation of foreground pixels in feature space to achieve the Fuzzy C-mean（FCM） clustering. It obtains feasible solutions of fitting equation and the distance between adjacent feasible solutions by using specific pixel fill fitting,image thinning and data fitting based on edge detection. It converts image pitch value to an actual pitch value by the image scale. Experimental results show that the algorithm can calculate the cable pitch quickly and efficiently, and the absolute error of cable pitch between the proposed algorithm and laser ranging algorithm is about 0. 75% . It has a high degree of measurement precision.

关 键 词：线缆节距 特征空间转换 模糊C均值聚类 边缘定位 数据拟合 激光测距 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]