摘要
憎水性图像水珠形状提取是绝缘子憎水性等级评价的重要前提。Snake模型在提取边界轮廓时存在着对初始轮廓敏感、边界信息捕捉能力较差等问题,为此笔者在S-L模型的基础上,利用水珠图像的先验信息来改进S-L模型对水珠形状的提取过程,在一定程度上克服了Snake模型的缺点,提高了轮廓形变过程中的拓扑能力,提高了形状提取的精确度。实验结果表明,该方法有效地实现了水珠形状的提取,为后续绝缘子性能的等级评价提供依据。
Water drops extraction of hydrophobic image is an important precondition to evaluate the hydrophobic performance of insulators. The problems associating with initialization and poor boundary catching ability limit the application of Snake model to water drops extraction. A new method based on S-L model is proposed for water drops extraction. The prior knowledge of water drops is used in confirming S-L model＇s energy function and the relevant weighting parameters. The water drops extraction based on the S-L model is improved by the adaptive energy item, significantly increasing the extraction precision, and overcoming the shortcomings of the original Snake model. Water drops extraction is improved with the proposed method to provide evidence for further grade evaluation of insulators.
出处
《高压电器》
CAS
CSCD
北大核心
2009年第6期37-39,共3页
High Voltage Apparatus
