基于Canny算子的复合绝缘子憎水性图像边缘检测

憎水性检测对于保障复合绝缘子安全可靠运行具有重要意义，在复合绝缘子憎水性自动识别过程中如何准确检测出水珠（或水迹）的边缘至关重要．为此，提出一种基于Canny算子和数学形态学的图像边缘检测方法，采用自适应局部灰度均衡对憎水性图像进行增强，减小因水珠透明性导致的目标与背景的相似度；采用形态学滤波，降低噪声影响、保持边缘细节；采用Ostu阈值法自适应的设定高低阈值．实验结果表明，该算法能够准确获取复合绝缘子表面水珠（或水迹）信息，为开展复合绝缘子憎水等级自动识别奠定了基础．

Hydrophobic image edge detection for composite insulator based on Canny operator

The hydrophobicity detection is important for safety operations of composite insulators. However, the key of automatic detection of composite insulators hydrophobic level is how to accurately detect the water drops edge （or water stains）. A method of edge detection based on Canny operator and Mathematical Morphology was presented in this paper. Firstly, the adaptive local gray balance was used for the enhancement of hydrophobicity image to decrease the similarity of the target and background due to the transparency of the drops. Then, the morphological filtering was adopted to reduce the impact of noises and keep the detail of edge. Finally, the Ostu threshold method was used to set thresholds. The results show that the algorithm can effectively extract the shape information of the water drops （or water stains） on composite insulator surface. It lays the foundation for automatic detection of composite insulators hvdrophobie level.