Skip to main content

Instance Segmentation on Jetson

This section demonstrates instance segmentation using Ultralytics YOLO segmentation models on NVIDIA Jetson. Examples include image, video, and real-time camera inference.

1. Enable Optimal Jetson Performance​

For best inference speed, enable maximum power and clocks.

Enable MAX Power Mode​

sudo nvpmodel -m 2

Enable Jetson Clocks​

sudo jetson_clocks

2. Instance Segmentation on Images​

Enter Demo Directory​

cd ~/ultralytics/ultralytics/yahboom_demo

Run Image Segmentation Script​

python3 02.segmentation_image.py

Results are saved to:

~/ultralytics/ultralytics/output/

Sample Code (Image Segmentation)​

from ultralytics import YOLO

model = YOLO("yolo11n-seg.pt")
results = model("assets/zidane.jpg")

for r in results:
    r.show()
    r.save(filename="output/zidane_output.jpg")

3. Instance Segmentation on Video​

Run Video Segmentation Script​

python3 02.segmentation_video.py

Output video location:

~/ultralytics/ultralytics/output/

Sample Code (Video Segmentation)​

import cv2
from ultralytics import YOLO

model = YOLO("yolo11n-seg.pt")
cap = cv2.VideoCapture("videos/people_animals.mp4")

width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
fps = int(cap.get(cv2.CAP_PROP_FPS))

out = cv2.VideoWriter(
    "output/people_animals_output.mp4",
    cv2.VideoWriter_fourcc(*"mp4v"),
    fps,
    (width, height)
)

while cap.isOpened():
    ret, frame = cap.read()
    if not ret:
        break

    results = model(frame)
    annotated = results[0].plot()
    out.write(annotated)

cap.release()
out.release()

4. Real-Time Instance Segmentation​

  • USB Camera: OpenCV VideoCapture(0)
  • CSI Camera: GStreamer pipeline (nvarguscamerasrc)

Real-time processing follows the same inference logic as video segmentation.

5. Notes​

  • Segmentation models output pixel-level masks
  • Suitable for defect contours and object separation
  • Use Nano segmentation models for real-time inference
  • Export to TensorRT for production deployment

Maintained by HemiHex for Jetson-based advanced vision workflows.