形态学边界增强算子及多界面检测中的应用

试管稠油油水界面准确检测对稠油破乳过程决策及操作优化具有重要作用。利用图像分析方法能够实时提取界面信息。由于稠油易附着玻璃试管,噪声严重,且不同界面明暗差异大,常规的图像界面提取方法难以同时提取不同界面。提出一种自适应边界检测算子,分别增强暗区域与亮区域。灰度形态学开运算,其运算结果分别膨胀和腐蚀处理,二者结果相减,增强暗区域。亮区域边界增强与暗区域边界增强是对偶运算。与传统的边界检测算子比较,该方法具有更好的抗噪性和定位能力。

Morphological boundary enhancement operator and multi-interface Detection

Accurate detection of viscous oil-water interface in the test tube would be important for the optimization of viscous oil demulsification. Interface information can be extracted by image analysis method in-time. Viscous oil is easy to stick on the glass tube. There is heavy noise in the tube. Different interfaces have different brightness. The conventional image interface extraction method is difficult to extract the different interfaces simultaneously. An adaptive boundary detection operator is proposed in this paper which can separately enhance the dark area and bright area. To enhance the dark area, the image is carried out by gray morphological open operation; the result is then processed by dilation and erosion operation respectively, the two results are then subtracted. Bright area boundary enhancement and dark area boundary enhancement are dual operation. The experiment can extract the interface of the viscous oil-water interface and the upper viscous oil-air interface. Compared with the conventional method, this method has more accurate and anti-noise performance.