摘要
针对多旋翼无人机(UAV)在电力巡检中的绝缘子跟踪问题,提出一种尺度自适应的绝缘子跟踪算法,采用稀疏投影的方式对原始图像特征进行降维,使用朴素贝叶斯分类器进行二分类,改进传统压缩感知(CS)跟踪搜索框固定问题,利用绝缘子的Lab空间特性进行分割,根据分割结果中绝缘子有效像素所占比例来改变搜索框的尺度,实现跟踪中的尺度自适应。实验结果表明:该算法能够在实验室和野外环境下自适应绝缘子尺度变化,对未来电力巡检智能化具有重大意义。
Aiming at insulator tracking problems of multi-rotor unmanned aerial vehicle （UAV） in power inspection, put forward a kind of dimension self-adaptive insulator tracking algorithm, adopting way of sparse projection to make dimensionality reduction on original image, it uses naive Bayesian classifier to make secondary classification to improve problem of classical compressive sensing （ CS ） tracking research frame, uses Lab space characteristics of insulator to make segmentation and changes scale of tracking box according to effective pixcl percentage of insulator to solve the fixed search box problem in Compressive Tracking. Experimental result indicates that this algorithm can make change in scale variations of insulator self- adaptively in lab and field environment, it has important meanings for intelligence of power inspection.
出处
《传感器与微系统》
CSCD
2016年第3期140-143,共4页
Transducer and Microsystem Technologies
关键词
压缩感知
投影矩阵
多尺度
绝缘子
跟踪
compressive sensing（CS）
projection matrix
multi-scale
insulator
tracking
