如何在Python的NumPy数组首行添加传感器数据表头？

Got it, let's sort out adding that header row to your IMU CSV file properly. The main issue with your current code is that it just appends data every time—we need to check if the file is brand new, and if so, write the header first before adding the sensor readings.

Here's the adjusted code with header support: #

import numpy as np
import time
import os  # Needed to check if the CSV file already exists

# Define your header to match the order of your sensor data
# Adjust these names if your sensor outputs have different dimensions/names
header = [
    "Timestamp",
    "Acceleration(g)_x", "Acceleration(g)_y", "Acceleration(g)_z",
    "Acceleration(angle)_x", "Acceleration(angle)_y", "Acceleration(angle)_z",
    "Gyroscope_raw_x", "Gyroscope_raw_y", "Gyroscope_raw_z",
    "Magnetometer_raw_x", "Magnetometer_raw_y", "Magnetometer_raw_z"
]

# Collect your IMU data in order
timestamp = time.time() * 1000  # Millisecond timestamp
imu_readings = []
imu_readings.extend(lsm6ds33.get_accelerometer_g_forces())
imu_readings.extend(lsm6ds33.get_accelerometer_angles())
imu_readings.extend(lsm6ds33.get_gyroscope_raw())
imu_readings.extend(lis3mdl.get_magnetometer_raw())

# Prepend timestamp to the readings to match your original logic
timestamped_row = np.array([timestamp] + imu_readings, dtype=np.float32)

csv_filename = "IMU_Readings_muh.csv"

with open(csv_filename, "ab") as f:
    # Write header only if the file doesn't exist yet
    if not os.path.exists(csv_filename):
        np.savetxt(f, [header], fmt="%s", delimiter=",")
    
    # Write the actual sensor data row
    np.savetxt(f, np.expand_dims(timestamped_row, axis=0), fmt="%4.8f", delimiter=",")

Key changes explained: #

• Added os module import to check if the CSV file already exists—this lets us only write the header once, on the first run.
• Defined a clear header list that matches the exact order of your data: timestamp first, followed by each sensor's axes in the order you collect them.
• Switched to using a Python list to collect readings first (instead of multiple np.append calls) — this is more efficient because numpy arrays are immutable, so each append creates a new array.
• Used fmt="%s" for the header to save the string labels correctly, while keeping your original %4.8f format for the numerical sensor data.

Quick note: #

Double-check that the number of header entries matches the number of data columns your code produces. For example, if get_accelerometer_g_forces() returns 2 values instead of 3, adjust the header to remove the extra axis entry.

内容的提问来源于stack exchange，提问作者rajy