| 基于聚合通道特征的防震锤锈蚀缺陷识别算法 |
投稿时间：2019-04-24 修订日期：2019-05-07  点此下载全文 |
| 引用本文： |
| 摘要点击次数: 288 |
| 全文下载次数: 0 |
|
| 基金项目:国家自然科学青年基金(61401001) |
|
| 中文摘要:高压输电线路中防震锤锈蚀会危害输电线路的安全运行。本文基于图像处理技术提出了一种基于聚合通道特征的防震锤检测和锈蚀缺陷识别的算法。该算法首先引入聚合通道特征（Aggregate Channel Features, ACF）分别提取无人机拍摄的输电线路图像中的颜色、梯度幅值和梯度方向直方图，构建多尺度ACF金字塔；利用滑窗法和Adaboost分类器检测图像中的防震锤，并使用非极大抑制操作得到最佳防震锤的位置；再结合Graph cuts算法实现防震锤图像的分割；最后采用RGB颜色模型识别防震锤锈蚀缺陷。实验结果表明该算法对防震锤位置的检测和锈蚀识别的精度较高。 |
| 中文关键词:聚合通道特征 滑窗法 Adaboost分类器 Graph cuts算法 |
| |
| Rust Defect Recognition Method for Anti-vibration Hammer Based on the Features of Aggregated Channels |
|
|
| Abstract:The anti-vibration hammer corrosion in high-voltage transmission lines will endanger the safe operation of transmission lines. In this paper, based on image processing technology, an algorithm for hammer detection and rust defect recognition based on aggregate channel features was proposed. The algorithm firstly introduced the Aggregate Channel Features (ACF) to extract the color, gradient amplitude and gradient direction histograms of the transmission line images were taken by the drone to construct a multi-scale ACF pyramids. The sliding window method and Adaboost classifier were used to detect the anti-vibration hammer in the image, and the optimal position of the hammer was obtained by non-maximum suppression operation. The image of the anti-vibration hammer was segmented by the Graph cuts algorithm. finally, the RGB color model was used to identify the anti-vibration hammer corrosion defect. The experimental results show that the algorithm has high precision in detecting the position of the anti-vibration hammer and rusting. |
| keywords:Aggregation channel feature sliding window method Adaboost classifier Graph cuts algorithm |
| 查看全文 查看/发表评论 下载pdf阅读器 |
位访问者,今日访问
