An Infusion Liquid Level Detection Method Based on Improved ROI Segmentation and Horizontal Lines Modification

In order to promote the development of intelligent medical system, researchers pay more attention to the intellectualization of intravenous infusion system. And the real-time liquid level detection is a key problem to achieve it. However, the existing researches only focus on the liquid level detection under clear and simple background. In practice, infusion bottle in complex background is more common, thus an infusion liquid level detection method based on improved region of interest (ROI) segmentation and horizontal lines modification is proposed to realize the liquid level detection for complicated environment. According to the features of the infusion bottle, the region of it can be segmented and the liquid level can be further detected. The experiments results demonstrate the performance to detect the liquid levels under clear and complex background, bright and dark light conditions, and the effectiveness of the proposed method is proved.