pylabrobot.brooks.precise_flex.PreciseFlexVision.capture_image#

async PreciseFlexVision.capture_image(camera: Literal['front', 'bottom'] | int = 'front') → ndarray#

Fetch one frame for camera directly off the vision server as an array, no motion.

Triggers a frame with system.cameraacquire and reads results off the vision server’s image stream until the matching Primary Image [n] arrives, discarding non-image and other-camera results, then decodes its JPEG to an RGB array. Requires a configured vision server (raises if none was set up). For a saved file on the vision server host instead, use save_image.

This grabs the camera’s current hardware state; it does NOT apply pending acquire-tool settings. Change one first with _set_camera_setting (which applies it) for it to show in the frame.

Parameters:

camera (Literal['front', 'bottom'] | int) – which gripper camera - "front"/1 (front-facing) or "bottom"/2 (downward).

Returns:

The full-resolution frame as an RGB numpy array (height x width x 3, uint8).

Raises:

  • TimeoutError – if no frame arrives off the vision server image stream before the read times out.

  • RuntimeError – if the image stream ends before the requested frame is seen.

Return type:

ndarray