Model prediction

1. Best performance mode

Enable Jetson Clocks

Enabling Jetson Clocks will ensure that all CPU and GPU cores run at maximum frequency:

sudo jetson_clocks

2. Model prediction

2.1. CLI usage

CLI currently only supports calling USB cameras. Nuwa camera users can directly modify the previous python code to call onnx and engine models!

xxxxxxxxxx
yolo predict model=best.engine source=0 save=False show # For multiple cameras, change the number after source.

2.2, Python usage

Use best.engine to predict the camera image:

Start the camera.

#usb camera
ros2 launch usb_cam camera.launch.py 
#nuwa camera
ros2 launch ascamera hp60c.launch.py

Enter the code folder:

xxxxxxxxxx
cd /home/jetson/ultralytics/ultralytics/yahboom_demo

Run the code: Click the preview image, press the q key to terminate the program!

xxxxxxxxxx
python3 06.orange_camera_usb.py

Effect preview

Yolo recognizes the output video location: /home/jetson/ultralytics/ultralytics/output/

Sample code:

​x
#!/usr/bin/env python3
import rclpy
from rclpy.node import Node
from sensor_msgs.msg import Image, CompressedImage
from cv_bridge import CvBridge
import cv2
from ultralytics import YOLO
import os
​
class Image_detection(Node):
    def __init__(self):
        super().__init__('Image_detection')
        self.model = YOLO("/home/jetson/ultralytics/ultralytics/data/yahboom_data/orange_data/runs/detect/train/weights/best.engine")
        self.camera_type = os.getenv('CAMERA_TYPE', 'usb')
        self.bridge = CvBridge()
        if self.camera_type == 'usb':
            topic_name = '/usb_cam/image_raw' 
        else:
            topic_name = '/ascamera_hp60c/camera_publisher/rgb0/image' 
​
        self.subscription = self.create_subscription(Image,topic_name, self.image_callback,10)
​
        # Get the video frame size and frame rate
        frame_width = 640
        frame_height = 480
        fps = 15
        
        output_path = "/home/jetson/ultralytics/ultralytics/output/06.orange_camera_usb.mp4"
        fourcc = cv2.VideoWriter_fourcc(*'mp4v')  # You can use 'XVID' or 'mp4v' depending on your platform
        self.out = cv2.VideoWriter(output_path, fourcc, fps, (frame_width, frame_height))
​
    def image_callback(self, msg):
        cv_image = self.bridge.imgmsg_to_cv2(msg, desired_encoding='bgr8')
 
        self.proecc(cv_image)
​
            
# Loop through the video frames
    def proecc(self,frame):
        # Run YOLO inference on the frame
        results = self.model(frame)
​
        # Visualize the results on the frame
        annotated_frame = results[0].plot()
​
        # Write the annotated frame to the output video file
        self.out.write(annotated_frame)
​
​
        # Display the annotated frame
        cv2.imshow("YOLO Inference", cv2.resize(annotated_frame, (640, 480)))
​
        # Break the loop if 'q' is pressed
        cv2.waitKey(1) & 0xFF == ord("q")
​
    def cancel(self):
        cv2.destroyAllWindows()
        self.out.release()
        
​
def main(args=None):
    rclpy.init(args=args)
    node = Image_detection()
    try:
        rclpy.spin(node)
    except KeyboardInterrupt:
        pass
    finally:
        node.cancel()
        node.destroy_node()
        rclpy.shutdown()
​
if __name__ == '__main__':
    main()