心率和血氧饱和度作为人体重要的生理参数，在个人健康管理、疾病早期识别和干预方面发挥着不可或缺的作用。基于成像光电容积描记（IPPG）的非接触式生理参数检测方法，不仅克服了传统接触式检测的限制，还让远程实时监测生理参数成为可能。尽管该方法在便捷性、经济性和易用性上具有优势，但采集设备成本高、面部运动产生的伪影难消除和实时性差等问题仍是挑战。针对这些挑战，本文利用手机摄像头在头部运动场景下对非接触式心率和血氧饱和度的准确性检测进行了研究与实验，主要工作如下：

（1）针对现有非接触式心率检测方法因仅提取绿色通道的信息作为IPPG信号而遗漏部分心率信息的问题，提出一种BRGB方法。将R-G与R+G-2×B作为基底向量对RGB通道信号进行组合，以生成周期性更为凸显的IPPG信号，进而提高心率检测的准确性。其次，针对变分模态分解的分解层数和惩罚因子参数需要主观选择的问题，引入全局搜索策略的鲸鱼优化算法对这些参数进行优化，以避免参数选择不当导致的过分解或欠分解现象。在VIPL-HR数据集上进行多组实验对比，结果表明，该算法得到的心率平均绝对误差为2.63bpm，符合中国人民共和国医药行业规定小于5bpm的误差标准。

（2）针对运动伪影下非接触式血氧饱和度检测准确性低的问题，提出一种改进的自适应噪声完全集合经验模态分解联合小波阈值的算法。通过应用该算法对蓝色通道和红色通道信号进行去噪，获取了高信噪比的脉搏波信号，进而提高了血氧饱和度的准确性。在VIPL-HR数据集和自采数据集上进行多组实验对比，结果显示，该算法得到的血氧饱和度平均绝对误差为1.12%，与其他算法相比具有较低的误差。

实验结果验证了本文所提出方法应用于非接触式心率与血氧饱和度检测的有效性，为该领域的实际应用提供了参考依据。同时，在上述研究和实验基础上，为实现生理参数的实时测量，本文使用Python编程语言搭建了一个能够实时显示人体心率及血氧饱和度数值的应用系统，用来提升用户体验的友好性和便利性。

Heart rate and blood oxygen saturation, as important physiological parameters of the human body, play an indispensable role in personal health management, early disease identification, and intervention. A non-contact physiological parameter detection method based on imaging photoelectric volumetric tracing (IPPG) technology not only overcomes the limitations of traditional contact detection but also makes it possible to monitor physiological parameters remotely and in real time. Despite the advantages of this method in terms of convenience, economy, and ease of use, the high cost of acquisition equipment, the difficulty of eliminating artifacts generated by facial motion, and the poor real-time performance are still challenges. To address these challenges, in this paper, we conducted research and experiments on the accurate detection of non-contact heart rate and oxygen saturation in head movement scenarios using a cell phone camera, and the main work is as follows:

(1)Aiming at the problem that existing non-contact heart rate detection methods miss part of the heart rate information due to extracting only the information of the green channel as the IPPG signal, a BRGB method is proposed. R-G and R+G-2×B are combined as base vectors on the RGB channel signals to generate IPPG signals with more prominent periodicity, which in turn improves the accuracy of heart rate detection. Second, to address the problem that the number of decomposition layers and the penalty factor parameters of the variational modal decomposition need to be subjectively selected, the whale optimization algorithm with a global search strategy is introduced to optimize these parameters to avoid the phenomenon of over-decomposition or under-decomposition caused by improper parameter selection. Multiple sets of experimental comparisons were performed on the VIPL-HR dataset, and the results showed that the algorithm yielded an average absolute error of heart rate of 2.63 bpm, which conformed to the error criterion of less than 5 bpm stipulated by the pharmaceutical industry of the People's Republic of China.

(2)Aiming at the low accuracy of non-contact blood oxygen saturation detection under motion artifacts, an improved algorithm based on adaptive noise complete set empirical mode decomposition combined with wavelet threshold is proposed. By using this algorithm to denoise blue channel and red channel signals, pulse wave signals with a high signal-to-noise ratio are obtained, and the accuracy of blood oxygen saturation is improved. Experiments on the VIPL-HR data set and self-collected data set show that the mean absolute error of blood oxygen saturation obtained by this algorithm is 1.12%, which has a lower error than other algorithms.

The experimental results verify the effectiveness of the proposed method in non-contact heart rate and blood oxygen saturation detection and provide a reference for practical application in this field. At the same time, based on the above research and experiments,to realize the real-time measurement of physiological parameters, this paper uses Python programming language to build an application system that can display the human heart rate and blood oxygen saturation values in real time, to improve the friendliness and convenience of user experience.