基于局部特征和空间顺序的航拍绝缘子检测算法研究与实现
毕业论文关键词：绝缘子检测；局部特征；空间顺序；特征匹配
Research and Implementation of Aerial Insulator Detection Algorithm Based on Local Feature and Spatial Order
Abstract：The detection of targets with complex backgrounds in aerial images is a challenging task. In recent years, the feature-based image matching technology has become a popular research direction for domestic and foreign scholars because of its excellent performance. This paper presents an aerial image insulator detection algorithm based on local feature and spatial order. Firstly, the local features in the aerial image of the aerial insulator are detected and represented by a multi-scale and multi-feature descriptor. Then, several spatial order features are obtained by training these local features. Finally, the feature matching strategy matches the insulator feature points and determines the area where the insulator is located. The experimental results show the accuracy and robustness of the proposed algorithm by detecting the insulators in different aerial insulators.
Key words: insulator detection； local feature； spatial orders； feature matching
目 录
摘要 1
关键词 1
Abstract 1
Key words 1
1绪论 2
1.1课题研究背景及意义 2
1.2国内外研究状况 2
1.2.1绝缘子检测研究现状 2
1.2.2相关技术现状 3
1.3本文的主要内容和结构安排 3
1.3.1本文的主要内容 3
1.3.2本文的结构安排 4
2航拍绝缘子图片的局部特征检测与描述 5
2.1改进的harris角点检测算法 5
2.2局部特征描述 6
2.2.1SPIN描述符及RIFT描述符概述 6
2.2.2多尺度多特征描述符MSMF 6
3绝缘子特征库生成（训练阶段） 9
3.1视觉词典（Lvv）的生成 9
3.2空间顺序特征（SOF）的生成 10
3.2.1空间顺序特征描述 10
3.2.2空间顺序特征（SOF）生成 10
4绝缘子的特征匹配及目标定位 12
4.1特征匹配 13
4.2绝缘子区域定位 14
4.2.1基于直线拟合的目标区域定位 14
4.2.2基于投票矩阵的目标区域定位 14
5实验设计与结果分析 14
5.1实验环境和实验设置 14
5.1.1实验环境 14
5.1.2实验设置 15
5.2实验设计 15
5.2.1特征提取与训练实验设计 15
5.2.2绝缘子识别与定位实验设计 17
5.3评价指标 18
5.4实验结果与分析 18
5.4.1特征提取与训练实验结果 18
5.4.2绝缘子识别定位实验结果分析 21