APP下载

多特征加权指数熵在光斑中心定位中的应用

2020-12-23张明富马月辉段梦琨郑军飞孙晓贤

现代电子技术 2020年23期

张明富 马月辉 段梦琨 郑军飞 孙晓贤

摘  要: 激光光斑中心检测是视觉测量中常用的关键技术,检测算法的优劣直接影响了测量的精度。针对现有光斑定位算法存在定位精度低、抗干扰能力差等问题,提出一种改进的激光光斑中心定位算法。该算法在Shannon熵的基础上引入指数熵,并根据光斑图像特征选取类内离散度最小和灰度对比度最大对指数熵进行加权处理,将使加权指数熵值最大时的灰度值作为阈值对光斑图像进行阈值分割,由此消除噪声干扰,最后由质心法进行光斑中心定位。实验表明,在兼顾稳定性的同时,与传统阈值分割方法相比,光斑中心定位误差降低了22%以上。

关键词: 视觉测量; 光斑定位; 指数熵应用; 阈值分割; 加权处理; 噪声干扰消除

中图分类号: TN911.73?34; TP391.41                文献标识码: A                 文章编号: 1004?373X(2020)23?0020?04

Abstract: Detection of laser spot center is a key technology commonly used in vision measurement. The quality of detection algorithm affects the accuracy of measurement directly. In allusion to the problems of low positioning accuracy and poor anti?jamming ability of the existing spot location algorithms, an improved laser spot center positioning algorithm is proposed. The exponential entropy is introduced on the basis of Shannon′s entropy. According to the characteristics of the light?spot image, the exponential entropy is weighted by choosing the smallest intra?class dispersion and the largest gray contrast ratio. The gray value of the maximum weighted exponential entropy is used as the threshold to segment the spot image, and eliminate noise interference. The center positioning of the light?spot has been realized by the centroid method. The experiment results show that, in comparison with the traditional threshold segmentation method, the error of the laser spot center positioning algorithm is reduced by more than 22% while it gives consideration of stability.

Keywords: vision measurement; light?spot positioning; exponential entropy application; threshold segmentation; weighted processing; noise jamming elimination

0  引  言

基于视觉的形变测量系统具有精度高、非接触、受环境影响小等优点,在桥梁、隧道等重点监测场合应用广泛[1?2]。而大多数视觉测量系统选取亮度集中、抗传输介质干扰能力强的光斑作为待测物[3],因此快速准确的光斑定位是实现精确形变测量的前提。

常用的光斑定位算法有插值法[4]、二维高斯拟合法[5?6]、灰度质心法[7]以及多种改进算法。插值法通常以光斑最亮点作为插值中心,选取插值中心[x]轴和[y]轴4个节点进行插值计算。插值法计算量小、运算速度快,但计算精度极易受所选插值节点灰度波动影响。对于边缘含强干扰亮点的光斑图,插值中心易误选。……

登录APP查看全文