I need help with YOLO v8

[8ussai]

i am trying to use the pose model with the object detection model to extract distance and other data between person and object using open cv

how can i do it if i can

[UltralyticsAssistant]

👋 Hello @8ussai, thanks for reaching out to Ultralytics 🚀! This is an automated response to help you get started quickly—an Ultralytics engineer will also assist you here soon.

We recommend a visit to the Docs for new users where you can find many Python and CLI usage examples and where many of the most common questions may already be answered.

If this is a 🐛 Bug Report, please provide a minimum reproducible example to help us debug it.

If this is a custom usage ❓ Question (e.g., combining pose and object detection with OpenCV), please share:
-Your ultralytics version and OS/GPU details
-The exact code snippet you’re using and steps to reproduce
-Example images/frames and a brief description of the expected vs. actual result
-Any logs or console output
Also, verify you are following our Tips for Best Training Results if training is involved.

Join the Ultralytics community where it suits you best. For real-time chat, head to Discord 🎧. Prefer in-depth discussions? Check out Discourse. Or dive into threads on our Subreddit to share knowledge with the community.

Upgrade

Upgrade to the latest ultralytics package including all requirements in a Python>=3.8 environment with PyTorch>=1.8 to verify your issue is not already resolved in the latest version:

pip install -U ultralytics

Environments

YOLO may be run in any of the following up-to-date verified environments (with all dependencies including CUDA/CUDNN, Python and PyTorch preinstalled):

• Notebooks with free GPU:
• Google Cloud Deep Learning VM. See GCP Quickstart Guide
• Amazon Deep Learning AMI. See AWS Quickstart Guide
• Docker Image. See Docker Quickstart Guide

Status

If this badge is green, all Ultralytics CI tests are currently passing. CI tests verify correct operation of all YOLO Modes and Tasks on macOS, Windows, and Ubuntu every 24 hours and on every commit.

[glenn-jocher]

You can run YOLO pose and detection on the same frame, then compute the pixel distance between a person (pose box or keypoints) and any detected object; minimal example:

import cv2, numpy as np
from ultralytics import YOLO

det_model = YOLO("yolo11n.pt")          # object detection
pose_model = YOLO("yolo11n-pose.pt")    # human pose

cap = cv2.VideoCapture(0)
while True:
    ok, frame = cap.read()
    if not ok:
        break

    det_res = det_model(frame, conf=0.25)     # optionally set classes=[...] to target a specific object
    pose_res = pose_model(frame, conf=0.25)

    boxes_obj = det_res[0].boxes
    boxes_ppl = pose_res[0].boxes
    kpts = pose_res[0].keypoints

    if len(boxes_obj) and len(boxes_ppl) and kpts is not None:
        o_xyxy = boxes_obj.xyxy.cpu().numpy()
        p_xyxy = boxes_ppl.xyxy.cpu().numpy()

        o_centers = np.column_stack(((o_xyxy[:,0]+o_xyxy[:,2])/2, (o_xyxy[:,1]+o_xyxy[:,3])/2))
        p_centers = np.column_stack(((p_xyxy[:,0]+p_xyxy[:,2])/2, (p_xyxy[:,1]+p_xyxy[:,3])/2))

        D = np.linalg.norm(p_centers[:, None, :] - o_centers[None, :, :], axis=-1)
        pi, oi = np.unravel_index(np.argmin(D), D.shape)
        d_px = float(D[pi, oi])

        c1, c2 = tuple(p_centers[pi].astype(int)), tuple(o_centers[oi].astype(int))
        cv2.line(frame, c1, c2, (0, 255, 0), 2)
        cv2.putText(frame, f"{d_px:.1f}px", c1, cv2.FONT_HERSHEY_SIMPLEX, 0.6, (0,255,0), 2)

    cv2.imshow("person-object distance", frame)
    if cv2.waitKey(1) == 27:
        break

cap.release()
cv2.destroyAllWindows()

For an interactive click-to-measure solution with tracking, see the guide on Distance Calculation using YOLO11 and the DistanceCalculation API reference; for true metric distances use a depth sensor or calibration—our ROS example shows how to combine segmentation with a depth image in the ROS quickstart (Depth example).