yolact_server

Recognize objects from an image using YolactEdge.

class yolact_server.YolactServer

Bases: object

YOLACT ROS Action server.

cv2smi(cv, encode)

Convert from OpenCV image to sensor_msgs/Image.

Parameters

  • cv (ndarray) – OpenCV image before conversion.

  • encode (str) – Conversion encode. mono8: grayscale image. mono16 16-bit grayscale image. bgr8: color image with blue-green-red color order. rgb8: color image with red-green-blue color order. bgra8: BGR color image with an alpha channel. rgba8: RGB color image with an alpha channel.

Returns

converted Image. If False is returned, the conversion has failed.

Return type

sensor_msgs/Image

delete()
delete_unnecessary_result(cv_rgb, cv_d, classes, scores, boxes, masks)

Delete unnecessary recognition results.

Using max distance (if depth image is available), ignore id, specific id, NMS.

Parameters

  • cv_rgb (ndarray) – RGB image.

  • cv_d (ndarray) – Depth image.

  • classes (list[str]) – Classes.

  • scores (list[float]) – Socres.

  • boxes (list[list[int]]) – Boxes.

  • masks (list[list[float]]) – Masks.

Returns

New lists with unnecessary results removed.

Return type

list[str], list[float], list[list[int]], list[list[float]]

draw_mask(cv_src, mask, color, alpha=0.5, only_contour=True)

Drawing mask information on the input image.

Parameters

  • cv_src (ndarray) – Input image.

  • mask (_type_) – Mask information.

  • color (_type_) – Color of mask.

  • alpha (float, optional) – Transparency of mask. Defaults to 0.5.

  • only_contour (bool, optional) – Drawing only contours. The mask drawing process is too slow, so True is RECOMMENDED. Defaults to True.

Returns

Output image after drawing.

Return type

ndarray

draw_result(cv_rgb, cv_d, classes, boxes, masks)

“Drawing result on the input image.

Parameters

  • cv_src (ndarray) – Input image.

  • classes (list[str]) – Classes.

  • boxes (list[list[int]]) – Boxes.

  • masks (list[list[float]]) – Masks.

Returns

Output image after drawing.

Return type

ndarray

get_new_image(time, is_d)

Get a new images.

Parameters

  • time (ROS time) – Get images acquired at a time newer than “time.”

  • is_d (bool) – Whether or not to get the depth image.

Returns

RGB and Depth image.

Return type

sensor_msgs/Image

main(goal)

Main function.

Parameters

goal (hma_yolact_action/YolactAction) – Input informations.

return_failure(message)

Return a failure.

Parameters

message (str) – Warning messages to be output.

smi2cv(smi, encode, is_depth=False)

Convert from sensor_msgs/Image to OpenCV image.

Parameters

  • smi (sensor_msgs/Image) – Image before conversion.

  • encode (str) – Conversion encode. mono8: grayscale image. mono16 16-bit grayscale image. bgr8: color image with blue-green-red color order. rgb8: color image with red-green-blue color order. bgra8: BGR color image with an alpha channel. rgba8: RGB color image with an alpha channel.

  • is_depth (bool, optional) – Whether to convert depth image. Defaults to False.

Returns

Converted OpenCV image. If False is returned, the conversion has failed.

Return type

ndarray

subf_rgb(smi_rgb)

Callback function for RGB image.

Parameters

smi_rgb (sensor_msg/Image) – RGB image.

subf_rgbd(smi_rgb, smi_d)

Callback function for RGB and Depth images.

Parameters

  • smi_rgb (sensor_msg/Image) – RGB image.

  • smi_d (sensor_msg/Image) – Depth image.