基于蚁群算法的绝缘子憎水性等级判别方法

Hydrophobic Detection Based on Ant Colony Algorithm

憎水性是衡量绝缘材料电性能的一个重要指标,也是确保绝缘子安全运行的重要保证,且对绝缘子憎水性等级进行判决的关键是准确地分割出水珠(或水迹)区域,因此将蚁群算法引入到憎水性图像处理中,该方法首先采用自适应局部灰度均衡对图像进行增强,以减小水珠的透明性导致的目标与背景的相似度;然后采用蚁群算法建立憎水性图像的知觉图表,利用图表信息实现对图像边缘信息的提取,从而获取水珠(或水迹)的轮廓信息。最后采用改进的形状因子法对处理后的图像进行憎水性等级的判别。实验结果表明,该算法能够较好地提取污秽绝缘子憎水性图像的水珠(或水迹)的轮廓信息,并能够通过合适的判别方法准确地判别出憎水性等级。

Hydrophobicity of insulator surfaces is an important criterion of electrical performance of insulating material, and guarantees to ensure that the insulator is well operated. The key problem of measuring hydrophobicity is to detect the accurate beads （or watermarks） regions. Therefore, the ant colony algorithm （ACA） was introduced to hydrophobic image processing. First, the image was enhanced by adaptive local gray-scale balance io reduce the similarity of objectives and background caused by the transparency of water. Then, the ant colony algorithm was used to establish the perceived charts of the hydrophobic image. The water （or water traces} of the contour information was obtained using the chart information to extract the edge information. Finally,the shape factor of picture was applied to classify hydrophobic image. Experimental results show that this algorithm is efficient to hydrophobic images processing and hydrophobic images are classified accurately.