视觉检测开源库-加载模型并进行检测

最新推荐文章于 2025-03-18 09:38:41 发布

原创

最新推荐文章于 2025-03-18 09:38:41 发布 · 359 阅读

0 ·

CC 4.0 BY-SA版权

文章标签：

#视觉检测 #人工智能 #计算机视觉

继续完善模型工具加载和话题订阅发布功能，完善后的代码如下：在 __init__方法中创建订阅和发布者

...
from ament_index_python.packages import get_package_share_directory
from vision_msgs.msg import Detection2DArray, ObjectHypothesisWithPose, Detection2D
from sensor_msgs.msg import Image
from cv_bridge import CvBridge
import cv2
import yolov5


class YOLOv5Ros2(Node):
    def __init__(self):
        ...
        # 加载模型和工具
        if self.model_path == "":
            package_share_directory = get_package_share_directory('yolov5_ros2')
            self.model_path = package_share_directory + "/config/yolov5n.pt"
        self.yolov5_model = yolov5.load(self.model_path, self.device)
        self.bridge = CvBridge()
        # 创建订阅发布者
        self.yolo_result_pub = self.create_publisher(
            Detection2DArray, "yolo_result", 10)
        if self.pub_result_image:
            self.result_img_pub = self.create_publisher(
                Image, "result_image", 10)
        self.image_sub = sel