Utility node for image capture

[kevinh42]

During testing, we would like to capture videos for later analysis or for validation. Videos can be compressed to save on storage space (preferably lossless), and images shouldn't be kept in-memory once written to the video.

Node should:

• Subscribe to RGB (8-bit 3-channel)/depth image (16-bit 1-channel) topics (see how CvBridge is used in https://github.com/ubc-subbots/triton/blob/master/src/triton_gazebo/triton_gazebo/bounding_box_image_saver.py to get an OpenCV format image from a message)
• Save images into video files (OpenCV's VideoWriter class should support 16-bit grayscale with Gstreamer backend https://docs.opencv.org/4.5.1/dd/d9e/classcv_1_1VideoWriter.html)
• Maybe at regular intervals a new video file is created so we have multiple smaller videos instead of one large one.
• Can be in Python if this makes it easier

Node parameters:

• RGB image topic
• Depth image topic
• Save location
• Interval for video segments