时序数据插值，对GPS采样数据进行线性插值

此篇为上一篇的接续，对时序数据进行插值

上篇文章 解析kml文件，提取经纬度信息存入csv
将kml里面的信息提取了出来，但是有些数据间隔太大，甚至几十秒才一个采样点，需求对这样的数据进行插值，形成每秒一个采样点的数据，代码较简单，主要应用了scipy库的 interpolate 方法，希望对有同样需求的小伙伴有所帮助。

import pandas as pd
import os
import time
from scipy import interpolate

FinalPath = './final/'
CsvPath = './CSV/'
files = os.listdir(CsvPath)

for file in files:
    start = time.time()

    name, _ = os.path.splitext(file)
    filepath = CsvPath + file
    # 读入数据，对时间进行处理
    print('dealing ' + file)

    df = pd.read_csv(filepath)
    # df.drop_duplicates(subset=['time'], keep='first', inplace=True)
    df['time_diff'] = df['time'].apply(lambda x: pd.to_datetime(x)).diff()
    df['time_diff'] = df['time_diff'].apply(lambda x: x.total_seconds())
    df['time_diff'].fillna(0, inplace=True)
    df['time_t'] = df['time_diff'].cumsum(axis=0)

    a = int(df['time_t'].values.max())
    b = int(df.shape[0])
    print(f'时间范围 {a} s')
    # print(f'原数据 {b} s')
    print(f'需要添加 {a - b + 1} s数据')
    # print(df.head())
    # print(b)

    # 计算需要插值的坐标点
    t = list(df['time_t'].values)
    # print(len(t))
    t_all = [i for i in range(a)]
    insert_t = list(set(t_all).difference(set(t)))
    print(len(insert_t))

    # 计算插值
    y_lon = df['Lon'].values
    y_lat = df['Lat'].values
    y_alt = df['altitude'].values

    Flinear_lon = interpolate.interp1d(t, y_lon, kind='linear')
    y_new_lon = Flinear_lon(insert_t)
    # y_new_lon = np.array(y_new_lon).tolist()
    y_new_lon = [round(yi, 6) for yi in y_new_lon]

    Flinear_lat = interpolate.interp1d(t, y_lat, kind='linear')
    y_new_lat = Flinear_lat(insert_t)
    # y_new_lat = np.array(y_new_lat).tolist()
    y_new_lat = [round(yi, 6) for yi in y_new_lat]

    Flinear_alt = interpolate.interp1d(t, y_alt, kind='linear')
    y_new_alt = Flinear_alt(insert_t)
    # y_new_alt = np.array(y_new_alt).tolist()
    y_new_alt = [round(yi, 1) for yi in y_new_alt]

    # 插值的表
    df_t = pd.DataFrame(insert_t, columns=['time_t'])
    df_lon = pd.DataFrame(y_new_lon, columns=['Lon'])
    df_lat = pd.DataFrame(y_new_lat, columns=['Lat'])
    df_alt = pd.DataFrame(y_new_alt, columns=['altitude'])

    df_insert = pd.concat([df_t, df_lon, df_lat, df_alt], axis=1)

    # 与原表合并，形成秒采数据
    df = pd.concat([df, df_insert], ignore_index=True, join='inner', axis=0)
    df.sort_values(by='time_t', inplace=True, ascending=True, ignore_index=True)
    # print(df)

    df.to_csv(FinalPath + name + '_interpolated.csv', index=False)
    # print('done at ' + time.strftime('%Y-%m-%d %H:%M:%S'))

    end = time.time()
    print('time cost: ' + str(round((end-start), 3)) + ' s')
    print('--' * 20)