A breathing rate detection method based on infrared thermal imaging video

Abstract With the continuous enrichment of material life, people’s concern for personal health is increasing, and respiratory rate, as an important indicator for assessing human health, is particularly important in medical monitoring. In this paper, we propose a method for respiratory rate analysis using thermal infrared imaging images, which implements face recognition, performs image tracking through HOG combined with VSM, and finally extracts the nose region containing only the respiration-related areas to obtain the average grey curve, which effectively reduces the noise of irrelevant pixel regions and improves the accuracy of the detection results. In addition, this paper innovatively combines machine learning and OTSU threshold segmentation methods to achieve real-time detection of respiration data through time series prediction and threshold adaptive dynamic adjustment. After simulation analysis, the system achieves the expected accuracy and real-time performance in the simulated environment. After practical testing, the system proves the effectiveness of the proposed method in clinical applications, provides a reliable monitoring tool for healthcare professionals, facilitates timely detection and treatment of patients’ respiratory abnormalities, and significantly improves the level and efficiency of medical monitoring.