vehicle_common.cv.tracking#

Attributes#

_NAMED_PAYLOAD_BOUNDS

_PAPER_MASK_KEYS

image

Functions#

_normalize_payload_color(→ str)

_largest_contour_centroid(...)

_legacy_focused_payload_mask(→ numpy.ndarray)

_resolved_payload_mask(→ numpy.ndarray)

Use the focused-DLZ classifier first, then fall back to HSV if needed.

find_payload(→ Optional[Tuple[int, int, bool]])

Detect the requested payload paper or return the DLZ centre as a fallback.

rotate_image(→ numpy.ndarray)

Rotate an image by the specified angle.

find_dlz(→ Optional[Tuple[int, int, numpy.ndarray | None]])

Detect payload in image using color thresholding.

compute_3d_vector(→ Tuple[float, float, float])

Convert pixel coordinates to a 3D direction vector.

Module Contents#

_NAMED_PAYLOAD_BOUNDS#
_PAPER_MASK_KEYS#
_normalize_payload_color(payload_color: str | None, lower_payload: numpy.ndarray | None, upper_payload: numpy.ndarray | None) → str#
_largest_contour_centroid(mask: numpy.ndarray) → tuple[numpy.ndarray | None, int | None, int | None]#
_legacy_focused_payload_mask(focused_bgr: numpy.ndarray, lower_payload: numpy.ndarray | None, upper_payload: numpy.ndarray | None) → numpy.ndarray#
_resolved_payload_mask(focused_bgr: numpy.ndarray, requested_color: str, lower_payload: numpy.ndarray | None, upper_payload: numpy.ndarray | None, paper_masks: dict[str, numpy.ndarray]) → numpy.ndarray#

Use the focused-DLZ classifier first, then fall back to HSV if needed.

find_payload(image: numpy.ndarray, lower_zone: numpy.ndarray, upper_zone: numpy.ndarray, lower_payload: numpy.ndarray | None, upper_payload: numpy.ndarray | None, uuid: str, debug: bool = False, save_vision: bool = False, payload_color: str | None = None) → Tuple[int, int, bool] | None#

Detect the requested payload paper or return the DLZ centre as a fallback.

Parameters:

  • image (np.ndarray) – Input BGR image.

  • lower_zone (np.ndarray) – Lower HSV threshold for zone marker.

  • upper_zone (np.ndarray) – Upper HSV threshold for zone marker.

  • lower_payload (np.ndarray | None) – Optional HSV fallback lower bound.

  • upper_payload (np.ndarray | None) – Optional HSV fallback upper bound.

  • debug (bool) – If True, return an image with visualizations.

  • save_vision (bool) – If True, save the visualization image.

  • payload_color (str | None) – Requested paper colour name.

Returns:

A tuple (cx, cy, dlz_empty) if detection is successful; otherwise, None.

Return type:

Optional[Tuple[int, int, bool]]

rotate_image(image: numpy.ndarray, angle: float) → numpy.ndarray#

Rotate an image by the specified angle.

Parameters:

  • image (np.ndarray) – The input image.

  • angle (float) – The rotation angle in degrees. Typically, use the negative of the camera yaw to compensate for rotation.

Returns:

The rotated image.

Return type:

np.ndarray

find_dlz(image: numpy.ndarray, lower_pink: numpy.ndarray, upper_pink: numpy.ndarray, lower_green: numpy.ndarray, upper_green: numpy.ndarray, debug: bool = False) → Tuple[int, int, numpy.ndarray | None] | None#

Detect payload in image using color thresholding.

Parameters:

  • image (np.ndarray) – Input BGR image.

  • lower_pink (np.ndarray) – Lower HSV threshold for pink marker.

  • upper_pink (np.ndarray) – Upper HSV threshold for pink marker.

  • lower_green (np.ndarray) – Lower HSV threshold for green payload.

  • upper_green (np.ndarray) – Upper HSV threshold for green payload.

Returns:

A tuple (cx, cy, visualization_image) if detection is successful; otherwise, None.

Return type:

Optional[Tuple[int, int, np.ndarray | None]]

compute_3d_vector(x: float, y: float, camera_info: numpy.ndarray, altitude: float, offset_x: float = 0, offset_y: float = 0, offset_z: float = 0) → Tuple[float, float, float]#

Convert pixel coordinates to a 3D direction vector.

image#