【摘要】 为了能够对医院调剂环境下的药品图像进行精准分割,进一步提高住院药房设备的自动化水平,本文提出了一种基于药品形状、颜色特性的改进区域生长算法。改进区域生长算法充分利用了传统区域生长算法实时性好和保持边界信息的优点,加入了药品尤其是对胶囊形状、颜色特性的考虑,弥补了传统区域生长算法对于胶囊药品同物双色特性没有很好适应的缺点,有效地解决了药品分割在胶囊药品以及粘连药品图像分割时药品不完整、边界不准确的问题,有助于医院调剂环境下药品图像中数量和种类的识别。更多还原

【Abstract】 In order to accurately segment the medicine images in the hospital dispensing environment and further improve the automation level of the inpatient pharmacy equipment, this paper proposes an improved area growth algorithm based on the shape and color characteristics of the medicine. The improved area growth algorithm takes full advantage of the traditional area growth algorithm’s good real-time performance and maintains the boundary information. It adds the consideration of the medicine,especially the shape and color characteristics of the capsule, and makes up for the shortcoming of traditional area growth algorithm that it has not good adaptation for the dual-color characteristics of the capsule medicine. It effectively solve the problems of incomplete drugs and inaccurate boundaries when segmenting the image of capsule drugs and adhesive drugs in drug segmentation,which is helpful for the identification of the number and type of drug images in the hospital dispensing environment.更多还原