Non-contact monitoring of vital signs such as respiration rate, heart rate, and blood pressure has consistently remained a key research focus. Traditional non-contact methods for vital sign monitoring, such as those utilizing camera, Wi-Fi, and radar, generally necessitate maintaining subject and sensor stability to mitigate motion-induced signal interference. However, this limitation significantly restricts its application of non-contact methods in real-world dynamic environments. To address this challenge, we propose a novel contactless Augmented Reality-integrated Vital Signs Sensing System (AR-VitalSense), which enables non-contact, accurate and consistent vital sign monitoring and visualization in dynamic scenarios. Three major procedures have been developed to complete the proposed AR-VitalSense. Firstly, we integrate an IR-UWB radar, as well as optical and infrared cameras, to acquire raw signals for vital signs such as heart rate, respiration rate, body temperature, and blood pressure. Secondly, we propose a two-stage jitter compensation algorithm to mitigate interference caused by device and human body motion, ensuring accurate extraction of vital signs. Thirdly, to evaluate our proposed AR-VitalSense with other state-of-the-art (SOTA) algorithms, a dataset with three different scenarios was collected from 10 participants. The total recording duration was 300 min. Comprehensive experiments are subsequently conducted on this dataset. Experimental results demonstrated mean absolute error (MAE) of 5.6 BPM for heart rate (HR), 2.2 BPM for respiration rate (RR), 17.8 mmHg for systolic blood pressure (SBP), 9.3 mmHg for diastolic blood pressure (DBP), and 0.22 ℃ for body temperature, significantly outperforming existing methods. To the best of our knowledge, AR-VitalSense is the first system that combines multiple sensors to achieve non-contact vital signs monitoring and display in real-world dynamic environments.