Regarding Mech-Vision, for obtaining image nodes from the camera and configuring virtual camera options

Can Mech-Vision be used to read virtual cameras from simulation software and parse the position of captured targets? Virtual cameras in the simulation software can export color and depth images.

Can it be understood that configuring the paths for color and depth images in the “Capture Images from Camera” Step in Mech-Vision allows for the reading of images, as illustrated in the figure below?

If possible, are there specific requirements for the formats of the color and depth images? After configuring the file paths, I encountered an error stating that the color image cannot be found.

When selecting the folder, make sure to navigate to the subfolder where the final images are saved.

You can confirm the format of the color image as .jpg and the depth image as .png by double-clicking the files inside.


Check and modify the paths, ensuring that there are no non-English characters in the file paths.

First, confirm whether these images are saved by the camera. Secondly, the “Capture Images from Camera” step has specific requirements for file names and formats in virtual mode. Color images should be named as rgb_image_xxxx.jpg, and depth images should be named as depth_image_xxx.png.