메뉴
Use for visualization nodes in dora. Triggers on: dora-rerun, opencv-plot, rerun, visualization, plot, display, image display, bounding box visualization, 3D visualization, depth visualization, 可视化, 显示, 绘图
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
| name | hub-visualization |
| description | Use for visualization nodes in dora. Triggers on: dora-rerun, opencv-plot, rerun, visualization, plot, display, image display, bounding box visualization, 3D visualization, depth visualization, 可视化, 显示, 绘图 |
| globs | ["**/dataflow.yml","**/dataflow.yaml"] |
| source | https://github.com/dora-rs/dora-hub |
Rerun and OpenCV visualization for images, 3D data, and telemetry
| Node | Install | Description |
|---|---|---|
| dora-rerun | pip install dora-rerun | Comprehensive visualization with Rerun |
| opencv-plot | pip install opencv-plot | Simple OpenCV image + bbox display |
Powerful visualization using Rerun with 12 supported primitives.
pip install dora-rerun
- id: rerun
build: pip install dora-rerun
path: dora-rerun
inputs:
camera_feed: camera/image
detections: yolo/bbox
depth_sensor: realsense/depth
env:
IMAGE_WIDTH: 640
IMAGE_HEIGHT: 480
RERUN_MEMORY_LIMIT: 25%
README: |
# My Visualization
Description shown in Rerun
All inputs require a primitive metadata field to specify visualization type.
inputs:
front_camera:
source: camera/image
metadata:
primitive: "image"
depth_sensor:
source: realsense/depth
metadata:
primitive: "depth"
focal: [600, 600]
camera_position: [0, 0, 0]
# Sender node includes primitive in metadata
node.send_output("image", image_data, {
"width": 640,
"height": 480,
"encoding": "bgr8",
"primitive": "image"
})
| Primitive | Data Type | Required Metadata |
|---|---|---|
| image | UInt8Array | width, height, encoding |
| depth | Float32Array | width, height |
| text | StringArray | - |
| boxes2d | StructArray/Float32Array | format: "xyxy" or "xywh" |
| boxes3d | Float32Array | format, solid, color |
| masks | UInt8Array | width, height |
| jointstate | Float32Array | (requires URDF config) |
| pose | Float32Array [x,y,z,qx,qy,qz,qw] | - |
| series | Float32Array | - |
| points3d | Float32Array | color, radii |
| points2d | Float32Array | - |
| lines3d | Float32Array | color, radius |
metadata = {
"primitive": "image",
"width": 640,
"height": 480,
"encoding": "bgr8" # bgr8, rgb8, jpeg, png, avif
}
metadata = {
"primitive": "depth",
"width": 640,
"height": 480,
"camera_position": [0, 0, 0], # [x, y, z]
"camera_orientation": [0, 0, 0, 1], # [qx, qy, qz, qw]
"focal": [600, 600], # [fx, fy]
"principal_point": [320, 240] # [cx, cy] optional
}
bbox = {
"bbox": np.array([x1,y1,x2,y2,...]).flatten(),
"conf": np.array([0.95,...]),
"labels": np.array(["person",...])
}
metadata = {"primitive": "boxes2d", "format": "xyxy"}
# center_half_size format (default)
boxes = np.array([cx,cy,cz, hx,hy,hz, ...])
metadata = {
"primitive": "boxes3d",
"format": "center_half_size", # or "center_size", "min_max"
"solid": False, # wireframe (default) or solid
"color": [255, 0, 0] # RGB
}
points = np.array([x1,y1,z1, x2,y2,z2, ...], dtype=np.float32)
metadata = {
"primitive": "points3d",
"color": [0, 255, 0],
"radii": [0.01, 0.01, ...]
}
# Line segments: pairs of xyz points
lines = np.array([x1,y1,z1, x2,y2,z2, ...], dtype=np.float32)
metadata = {
"primitive": "lines3d",
"color": [0, 0, 255],
"radius": 0.005
}
# 7 values: position + quaternion
pose = np.array([x, y, z, qx, qy, qz, qw], dtype=np.float32)
metadata = {"primitive": "pose"}
- id: rerun
path: dora-rerun
inputs:
jointstate_robot: arm/joint_state
env:
robot_urdf: /path/to/robot.urdf
robot_transform: "0 0.3 0" # x y z offset
nodes:
- id: camera
build: pip install opencv-video-capture
path: opencv-video-capture
inputs:
tick: dora/timer/millis/33
outputs:
- image
- id: yolo
build: pip install dora-yolo
path: dora-yolo
inputs:
image: camera/image
outputs:
- bbox
- id: vlm
build: pip install dora-qwen2-5-vl
path: dora-qwen2-5-vl
inputs:
image:
source: camera/image
queue_size: 1
outputs:
- text
env:
DEFAULT_QUESTION: "Describe the scene."
- id: rerun
build: pip install dora-rerun
path: dora-rerun
inputs:
camera:
source: camera/image
metadata:
primitive: "image"
detections:
source: yolo/bbox
metadata:
primitive: "boxes2d"
description:
source: vlm/text
metadata:
primitive: "text"
env:
IMAGE_WIDTH: 640
IMAGE_HEIGHT: 480
RERUN_MEMORY_LIMIT: 25%
Simple OpenCV visualization with image, bounding boxes, and text overlay.
- id: plot
build: pip install opencv-plot
path: opencv-plot
inputs:
image: camera/image
bbox: yolo/bbox
text: vlm/text
env:
PLOT_WIDTH: 640 # Optional, defaults to image width
PLOT_HEIGHT: 480 # Optional, defaults to image height
image: UInt8Array
metadata = {"width": 640, "height": 480, "encoding": "bgr8"}
bbox: StructArray
bbox = {
"bbox": np.array([x1,y1,x2,y2,...]),
"conf": np.array([0.95,...]),
"labels": np.array(["person",...])
}
metadata = {"format": "xyxy"}
text: StringArray
text = pa.array(["Detection: person 95%"])
nodes:
- id: camera
build: pip install opencv-video-capture
path: opencv-video-capture
inputs:
tick: dora/timer/millis/33
outputs:
- image
- id: yolo
build: pip install dora-yolo
path: dora-yolo
inputs:
image: camera/image
outputs:
- bbox
- id: plot
build: pip install opencv-plot
path: opencv-plot
inputs:
image: camera/image
bbox: yolo/bbox
inputs:
image: camera/image # Primitive inferred from "image" in name
inputs:
front_camera:
source: camera/image
metadata:
primitive: "image"
| Variable | Description |
|---|---|
| IMAGE_WIDTH | Default image width |
| IMAGE_HEIGHT | Default image height |
| RERUN_MEMORY_LIMIT | Memory limit (e.g., "25%", "1GB") |
| README | Markdown text shown in Rerun |
| *_urdf | Path to URDF file for joint visualization |
| *_transform | Position offset "x y z" for URDF |
Use for dora CLI command questions. Triggers on: dora run, dora build, dora start, dora stop, dora new, dora up, dora destroy, dora list, dora ps, dora logs, dora check, dora graph, dora daemon, dora coordinator, CLI命令, 命令行, 运行数据流, 构建, 启动, 停止
Use for data recording and replay questions with dora. Triggers on: recording, replay, lerobot, data collection, dataset, rosbag, record data, play back, training data, HDF5, parquet, 数据记录, 数据回放, 数据收集, 训练数据
Use for dora-rs dataflow YAML configuration questions. Triggers on: dataflow.yml, dataflow.yaml, nodes:, inputs:, outputs:, timer, dora/timer, queue_size, env:, build:, path:, git:, branch:, tag:, restart_policy, send_stdout_as, operator:, operators:, YAML配置, 数据流配置, 节点配置, 输入输出, 定时器
Use for audio processing pipeline questions with dora. Triggers on: speech, TTS, STT, voice, audio, Whisper, Kokoro, text-to-speech, speech-to-text, voice activity, VAD, microphone, 语音, 音频, 文本转语音, 语音转文本, 语音识别
Use for robot control pipeline questions with dora. Triggers on: robot, arm, chassis, actuator, servo, motor, gripper, kinematics, FK, IK, URDF, joint, position control, velocity control, serial port, 机器人, 机械臂, 底盘, 执行器, 舵机, 电机, 运动学
Use for ML/Vision pipeline questions with dora. Triggers on: YOLO, detection, object detection, segmentation, tracking, VLM, SAM, camera, webcam, image processing, bbox, bounding box, rerun, visualization, 视觉, 目标检测, 分割, 跟踪, 摄像头, 图像处理