基于机器视觉及深度学习的静脉药物调配机器人药瓶识别

静脉药物调配机器人需要对数百种不同类型的药物进行调配,不同药物药瓶类型的识别、测量是精准调配的关键。对静脉药物调配机器人在药瓶类型分类、药瓶关键尺寸测量等不同任务要求情况下,运用基于机器视觉方法和深度学习框架YOLOv5s,实现了一种高效稳定的识别测量方法。结合机器视觉方法的测量准确性,深度学习方法对背景、光源的不敏感性,实现配药过程中不同环节药瓶的准确识别、测量。并通过实验定量地检测背景、环境光源对于识别结果的影响。

Drug Bottle Recognition of Intravenous Drug Dispensing Robot Based on Machine Vision and Deep Learning

Intravenous drug dispensing robots have to dispense up to hundreds of different types of drugs.The identification and measurement of different types of drug bottles is the key to precise deployment.Under the condition of different task requirements such as bottle type classification and bottle key size measurement,an efficient and stable recognition and measurement method was realized by using the method based on machine vision and deep learning framework YOLOv5s.Combined with the measurement accuracy of machine vision method and the insensitivity of deep learning method to background and light source,accurate identification and measurement of medicine bottles in different links in the dispensing process could be realized.The influence of background and ambient light source on the identification results was quantitatively detected through experiments.