Accessing camera + sensors

[Ohmagar]

Hi there,

is there a way to access the camera of an IPad? Or the sensors? I know some other python IDEs offer a module to access the gyroscope, the magnetometer and accelerometer (omz-software „The motion module allows you to access your iOS device’s motion sensor data (accelerometer, gyroscope, magnetometer).“).

Did/Do I just happen to miss it in Carnets or did you integrate ways to do that? Thank you so much for a reply in advance.

All the best, Ohm

[holzschu]

The camera is accessible through regular packages like OpenCV. I’ve seen people access the sensors using a bridge between Python and objective-C.

[Ohmagar]

Thank you for the quick response!

Unfortunately I ran into another problem:

error Traceback (most recent call last) Cell In [2], line 36 33 print("No available video devices found.") 35 if name == "main": ---> 36 main()

Cell In [2], line 31, in main() 28 print("Available video devices:", available_devices) 30 if available_devices: ---> 31 capture_and_display_frames(available_devices) 32 else: 33 print("No available video devices found.")

Cell In [2], line 15, in capture_and_display_frames(devices) 12 cap.release() 14 for idx, frame in enumerate(captured_frames): ---> 15 cv2.imshow(f"Captured Frame {idx}", frame) 17 cv2.waitKey(0) 18 cv2.destroyAllWindows()

error: OpenCV(4.5.2) /Users/holzschu/src/Xcode_iPad/Carnets/cpython/packages/opencv-python/opencv/modules/highgui/src/window.cpp:679: error: (-2:Unspecified error) The function is not implemented. Rebuild the library with Windows, GTK+ 2.x or Cocoa support. If you are on Ubuntu or Debian, install libgtk2.0-dev and pkg-config, then re-run cmake or configure script in function 'cvShowImage'

do you have an idea what the problem might be? A quick reinstall or is it something that’s rooted deeper? best regards

[holzschu]

Debugging is always easier if you provide the code along with the error message.

Working with the error message, I think that OpenCV is trying to open a new window to show the captured frames. That makes sense if you are running a Python script in a terminal, but less sense in a Jupyter notebook, where the image would be better as embedded inside the notebook. Inside iOS, the notion of "open a new window" does not exist, so this won't work.

This short script, for example, captures a frame from the front camera, and displays it inside the notebook:

import cv2 as cv
from matplotlib import pyplot as plt
cap = cv.VideoCapture(1)
ret, frame = cap.read()
plt.imshow(frame)
plt.show()

(VideoCapture(0) would give you the back camera).