6.4.10 YOLO-World Object Detection

Last Version: 12/09/2025

This guide demonstrates how to use the YOLO-World model for Open-Vocabulary object detection on SpacemiT computing hardware. You'll learn to run inference on both images and video streams, visualize the results, and process the detection data through ROS 2 topics.

YOLO-World Model Overview

YOLO-World is an open-vocabulary, zero-shot object detection model developed by Tencent AI Lab. First introduced in early 2024, it combines:

• The fast and efficient YOLO architecture (based on Ultralytics YOLOv8)
• Vision-language fusion technology
• Ability to detect objects described by arbitrary text prompts without category-specific training

Environment Setup

Install Dependencies

sudo apt install python3-venv python3-pip ros-humble-camera-info-manager \
ros-humble-image-transport python3-spacemit-ort

Source ROS2 Environment

source /opt/bros/humble/setup.bash

Set Up Python Virtual Environment

python3 -m venv ~/test3
source ~/test3/bin/activate
pip install -r /opt/bros/humble/share/jobot_yolo_world/data/requirements.txt

Update the Python path

export PYTHONPATH="$HOME/test3/lib/python3.12/site-packages":$PYTHONPATH

Image-Based Object Detection

Prepare a Test Image

cp /opt/bros/humble/share/jobot_yolo_world/data/test2.jpg .

Run Inference and Save Results Locally

ros2 launch br_perception yoloworld_infer_img.launch.py \
  img_path:='./test2.jpg' \
  class_names:="[fan, box]"

The detection results are saved as yoloworld_result.jpg in the current directory.

Terminal output shows detection details:

(test3) bianbu@bianbu:~$ ros2 launch br_perception yoloworld_infer_img.launch.py   img_path:='./test2.jpg' class_names:="[fan, box]"
[INFO] [launch]: All log files can be found below /home/bianbu/.ros/log/2025-08-13-16-09-04-608187-bianbu-217760
[INFO] [launch]: Default logging verbosity is set to INFO
[INFO] [yoloworld_img_node-1]: process started with pid [217761]
[yoloworld_img_node-1] /home/bianbu/test3/lib/python3.12/site-packages/clip/clip.py:6: UserWarning: pkg_resources is deprecated as an API. See https://setuptools.pypa.io/en/latest/pkg_resources.html. The pkg_resources package is slated for removal as early as 2025-11-30. Refrain from using this package or pin to Setuptools<81.
[yoloworld_img_node-1]   from pkg_resources import packaging
[yoloworld_img_node-1] All model files already exist and do not need to be downloaded
[yoloworld_img_node-1] conf_threshold: 0.2, iou_threshold: 0.45, class_names: ['fan', 'box']
[yoloworld_img_node-1] Init Model ..................
[yoloworld_img_node-1] all time cost:0.697995662689209
[yoloworld_img_node-1] x_min:980, y_min:469, width:466, height:424, label:box, confidence:0.72
[yoloworld_img_node-1] x_min:417, y_min:510, width:294, height:493, label:fan, confidence:0.29
[yoloworld_img_node-1] The object detection results are saved in: det_result.jpg
[INFO] [yoloworld_img_node-1]: process has finished cleanly [pid 217761]

Web-Based Visualization

Terminal 1 - Start inference node:

ros2 launch br_perception yoloworld_infer_img.launch.py \
  publish_result_img:=true \
  img_path:='./test2.jpg' \
  class_names:="[fan, box]"

Terminal 2 - Launch web visualization:

ros2 launch br_visualization websocket_cpp.launch.py image_topic:='/result_img'

The terminal will print the message as:

...
Please visit in your browser: http://<IP>:8080
...

Open a browser and visit the URL shown in the terminal (e.g., http://<IP>:8080) to view results:

Subscribe to Results

To view the inference results by

ros2 topic echo /inference_result

Or use this Python script to subscribe to inference results:

from rclpy.node import Node
from std_msgs.msg import Header
from jobot_ai_msgs.msg import DetectionResultArray, DetectionResult
import rclpy

class DetectionSubscriber(Node):
    def __init__(self):
        super().__init__('detection_sub')
        self.subscription = self.create_subscription(
            DetectionResultArray,
            '/inference_result',
            self.listener_callback,
            10)

    def listener_callback(self, msg: DetectionResultArray):
        self.get_logger().info(f"Frame: {msg.header.frame_id}")
        for det in msg.results:
            self.get_logger().info(
                f"[{det.label}] ({det.x_min},{det.y_min}) "
                f"{det.width}x{det.height} conf={det.conf:.2f}"
            )

def main(args=None):
    rclpy.init(args=args)
    node = DetectionSubscriber()
    rclpy.spin(node)
    node.destroy_node()
    rclpy.shutdown()

main()

Parameters Description

yoloworld_infer_img.launch.py Parameters

Parameter	Description	Default Value
img_path	Path to the input image for detection	data/detection/test.jpg
publish_result_img	Whether to publish detection results as an image message	false
result_img_topic	Topic name for published image results (if publish_result_img = true)	/result_img
result_topic	Topic name for published detection results	/inference_result
conf_threshold	Minimum confidence score for detections (filters low-confidence boxes)	0.2
iou_threshold	Threshold for removing overlapping detections (handles overlapping boxes)	0.45
class_names	Target objects to detect (can be categories or natural language descriptions)	["people"]

Video Stream Inference

Start USB Camera

ros2 launch br_sensors usb_cam.launch.py video_device:="/dev/video20"

Run Inference & Publish results

Terminal 1 - Start inference:

ros2 launch br_perception yoloworld_infer_video.launch.py \
  sub_image_topic:='/image_raw' \
  publish_result_img:=true \
  result_topic:='/inference_result' \
  class_names:="[box]"

Terminal 2 - Launch web visualization:

ros2 launch br_visualization websocket_cpp.launch.py image_topic:='/result_img'

The terminal will print the message as:

...
Please visit in your browser: http://<IP>:8080
...

Open a browser and visit the URL shown in the terminal (e.g., http://<IP>:8080) to view results.

Data-Only Inference (No Visualization)

If you only want to obtain the model inference results, run the following command:

ros2 launch br_perception yoloworld_infer_video.launch.py \
  sub_image_topic:='/image_raw' \
  publish_result_img:=false \
  result_topic:='/inference_result' \
  class_names:="[box]"

Subscribe to Results

To view the inference results by

ros2 topic echo /inference_result

Or use this Python script to subscribe to inference results:

from rclpy.node import Node
from std_msgs.msg import Header
from jobot_ai_msgs.msg import DetectionResultArray, DetectionResult
import rclpy

class DetectionSubscriber(Node):
    def __init__(self):
        super().__init__('detection_sub')
        self.subscription = self.create_subscription(
            DetectionResultArray,
            '/inference_result',
            self.listener_callback,
            10)

    def listener_callback(self, msg: DetectionResultArray):
        self.get_logger().info(f"Frame: {msg.header.frame_id}")
        for det in msg.results:
            self.get_logger().info(
                f"[{det.label}] ({det.x_min},{det.y_min}) "
                f"{det.width}x{det.height} conf={det.conf:.2f}"
            )

def main(args=None):
    rclpy.init(args=args)
    node = DetectionSubscriber()
    rclpy.spin(node)
    node.destroy_node()
    rclpy.shutdown()

main()

Parameters Descriptions

yoloworld_infer_video.launch.py Parameters

Parameter	Description	Default Value
sub_image_topic	Topic name for the input image	/image_raw
publish_result_img	Whether to publish detection results as an image message	false
result_img_topic	Topic name for published image results (if publish_result_img = true)	/result_img
result_topic	Topic name for published detection results	/inference_result
conf_threshold	Minimum confidence score for detections (filters low-confidence boxes)	0.2
iou_threshold	Threshold for removing overlapping detections (handles overlapping boxes)	0.45
class_names	Target objects to detect (can be categories or natural language descriptions)	["people"]