Non-contact heart rate measurement using low-cost RGB camera under complex light conditions

A non-contact heart rate measurement method using low-cost RGB video is proposed in this study. Only an RGB video of a human wrist is required as input for the proposed method. The method consists of spatial processing, gray value compensation, temporal processing, and heart rate measurement. Spatial processing runs firstly for preliminary image noise reduction. Gray value compensation is then used to compensate for the impact of ambient light changes. Temporal processing is used for pixel-by-pixel bandpass filtering. The filtered data is used at last for heart rate measurement. Experiments under various camera-wrist distances and light intensities are conducted to verify the effectiveness of the proposed method. Results show that the average measurement accuracy improves by 2.2 % and 1.9 % respectively by the gray value compensation algorithm under different camera-wrist distances and light intensities. High measurement accuracy can be obtained when the camera-wrist distance is within 10 cm and the light intensity is more than 400 lux.