1

I try to use python to read webots camera image, and publish it by ros topic. But when I write like: Img = camera.getImageArray() in the main loop, the simulation world will be very slow, speed like 0.1x. my program like this:

from controller import Robot, Camera
robot = Robot()
timestep = int(robot.getBasicTimeStep())
camera = Camera('camera')
camera.enable(100)
while robot.setp(timestep) != -1:
    img = camera.getImageArray()
    pass

If I delete this code "img = camera.getImageArray()" from main loop, then the simulation speed is increase to nomal speed. My webots version is Webots R2020b-rev1, os is Windows10 64-bit.I also try to run this code in Ubuntu16.04lts, it's same. Could anybody help me to this problem?

ke li
  • 11
  • 1

1 Answers1

0

You should use the getImage() method instead of getImageArray(). The documentation says "Another way to get the camera image in Python is to use the getImageArray method which returns a list<list<list>>. This three dimensional list can be directly used for accessing the RGB pixels value. However, this method runs significantly slower than the getImage method."

Also, the latest version of Webots is R2021a and may run faster than R2020b-rev1.

Olivier Michel
  • 755
  • 6
  • 10
  • thanks a lot, I try to use getImage() and the speed increase to 1.0x. I try to write a extra controller, pub the camera image to ros. So now I should convert the image list to opencvformat. – ke li Dec 21 '20 at 12:39