航拍宽幅图像的玻璃绝缘子定位研究

针对绝缘子爆裂和缺失将导致整个输电线路绝缘能力下降的常见故障，在宽幅航拍图像如何快速实现玻璃绝缘子的定位是绝缘故障检测的关键。提出一种通过机器学习的绝缘子定位算法。输电线路玻璃绝缘子串，利用Retinex算法与高斯滤波及形态学滤波对图像进行预处理和粗定位，建立不同样本数据提取特征，通过将二分类问题转化为多分类问题的集成学习方式，克服数据不平衡的问题，实现精准绝缘子定位，为进一步故障检测奠定了良好的基础。

Study on aerial image positioning of wide glass insulator

According to burst and lack of insulator will lead to the transmission line insulation capacity decreased the common fault,in wide aerial image to realize how fast the positioning of the glass insulator is insulation fault detection is the key.A kind of insulator location algorithm based on machine learning is proposed.Transmission line glass insulator strings,the Retinex algorithm and Gaussian filter and morphological filter for image preprocessing and coarse positioning,the establishment of different sample data feature extraction,the binary classification problem into multi class classification problem of ensemble learning,overcome the problem of unbalanced data, achieve precise positioning insulator,for further fault detection laid the good foundation.