摘要
为提高测井曲线自动跟踪技术的准确率,提出一种局部搜索和整体择优相结合、辅以曲线预测的实线自动跟踪算法。采用局部搜索来跟踪未受干扰的曲线,同时预测即将出现的曲线特征,以检验下一步的跟踪结果的正确性,并提出一种基于加权投票机制的整体择优算法以便从干扰所引起的多个分支中选择合适的分支。实验表明算法识别效果较好,速度快,且抗干扰性能较好,尤其是对于带有坐标网格的图纸,无需对图像进行去网格等预处理,可以节约预处理时间。
To improve the accuracy of automatically tracing of well log curves, an automatically tracing method for solid curves is proposed. It combines micro-searching method with macro-choosing method and is assisted by curve forecast. The micro-search is used to trace those solid curves which are not disturbed by other curves or coordinate grids, and the curve forecast is done in order to check the accuracy of the coming tracing points. A macro-choosing method based on weighted voling is proposed to choose the best suitable branch from the branches caused by disturbance. The experiments shows that the methods above are more accurate with fast speed, better anti-disturbance ability, and can save time because pre-processing step of deleting coordinate grids is not needed.
出处
《成都理工大学学报（自然科学版）》
CAS
CSCD
北大核心
2008年第2期177-180,共4页
Journal of Chengdu University of Technology: Science & Technology Edition
基金
四川省科技攻关项目(05GG021-026-03)
关键词
曲线自动识别
局部搜索
整体择优
加权投票
automatic curve tracing
micro-searching method
macro-choosing method
weighted votlng
