code/physiological_indicators.py

from utils_sig import *
import joblib
import numpy as np
from lightgbm import LGBMRegressor

import heartpy as hp
import scipy.signal as sig

class PhysiologicalIndicators:
    def __init__(self):
        self.heart_rate = 0
        self.respiratory_rate = 0
        self.heart_rate_variability = 0
        self.SpO2 = 0
        self.blood_pressure = 0
    
    def calculate_heart_rate(self, ippg_data, fps):
        # 计算心率的代码
        print("HR processing")
        self.heart_rate, self.respiratory_rate = hr_fft_2(ippg_data,  fps)
        # ippg = butter_bandpass(ippg_data, lowcut=0.6, highcut=4, fs=fps)
        # self.heart_rate, psd_y, psd_x = hr_fft(ippg, fs=fps)
        print("HR done")
        return self.heart_rate, self.respiratory_rate

    def calculate_heart_rate_variability(self, ippg_data, fps):
        # 计算心率变异性的代码
        # TODO: 实现心率变异性计算
        self.heart_rate_variability = calculate_hrv(ippg_data, fps)
        return self.heart_rate_variability

    def calculate_SpO2(self, ROI_list, ROI2_list):
        # 计算血氧饱和度的代码
        # TODO: 实现血氧饱和度计算
        ROI1_SpO2 = RGB_SpO2(ROI_list)
        ROI2_SpO2 = RGB_SpO2(ROI2_list)
        self.SpO2 = (ROI1_SpO2 + ROI2_SpO2) / 2
        return self.SpO2

    def calculate_blood_pressure(self, ippg_data):
        # 计算血压的代码
        ippg_data = np.array(ippg_data).reshape(len(ippg_data),1)

        bp_pred = []
        model_list = joblib.load( './code/model_weight/lgb_model_ppg2bp.pkl')
        for model in model_list:
            result = model.predict(ippg_data)
            bp_pred.append(result+10)
        bp_list = np.mean(bp_pred, axis=0)
        return bp_list,np.max(bp_list),np.min(bp_list)-15

    def calculate_HR(self, ROI_list, ROI2_list):
        # 计算HR的代码
        ROI1_HR = RGB_HR(ROI_list)
        ROI2_HR = RGB_HR(ROI2_list)
        print("ROI1_HR",ROI1_HR,"ROI2_HR",ROI2_HR)
        HR = (ROI1_HR + ROI2_HR) / 2
        return HR


    # 定义一个函数来计算心率
    def calculate_heart_rate_2(self,ppg_signal, sampling_rate):
        # 使用巴特沃斯滤波器处理信号，去除噪声
        nyquist_frequency = sampling_rate / 2.0
        low_cutoff_frequency = 0.5
        high_cutoff_frequency = 5.0
        filter_order = 2

        b, a = sig.butter(filter_order, [low_cutoff_frequency/nyquist_frequency, high_cutoff_frequency/nyquist_frequency], btype='band')
        filtered_signal = sig.filtfilt(b, a, ppg_signal)

        # 计算心率
        window_length = int(sampling_rate * 0.75)
        step_size = int(sampling_rate * 0.1)
        threshold = 0.4

        # 使用峰值检测算法来找到脉冲峰值
        peak_indexes, _ = sig.find_peaks(filtered_signal, distance=10)
        print("============================",peak_indexes,sampling_rate)
        # 计算时间间隔并计算心率
        # time_intervals = np.diff(peak_indexes) / float(sampling_rate)
        time_intervals = np.diff(peak_indexes) * 0.045
        heart_rate = 60.0 / np.mean(time_intervals)

        return heart_rate
    
    # 定义一个函数来从rppg信号中计算心率 
    def calculate_heart_rate_3(self,signal):
        wd, m = hp.process(signal, sample_rate = 100.0)

        return wd, m
        


    # def calculate_SPO2(self, ippg_chanel_data):
    #     # 计算血氧的代码
    #     ippg_chanel_data = np.array(ippg_chanel_data).reshape(len(ippg_chanel_data),6)

    #     SPO2_pred = []
    #     model_list = joblib.load( './model_weight/lgb_model_threechanel2spo2.pkl')
    #     for model in model_list:
    #         result = model.predict(ippg_chanel_data)
    #         SPO2_pred.append(result)
    #     SPO2_list = np.mean(SPO2_pred, axis=0)
    #     return SPO2_list