YOLO for Real-Time Object Detection: Complete End-to-End Guide (2026)

Beautiful real-time detection with bounding boxes and confidence labels 🚀

Assalam-o-Alaikum! Welcome to this complete hands-on guide. In this tutorial (and the video above), we cover YOLO26 – everything from training custom models to real-time inference on webcam/video with clean annotations using Ultralytics and Supervision.

We will go through:

• ✅ 1. Installation
• ✅ 2. Dataset preparation & annotation
• ✅ 3. Training a custom model
• ✅ 4. Real-time inference script (CLI app)
• ✅ 5. Running on webcam/video

1. Installation

pip install ultralytics supervision opencv-python typer

2. Dataset Preparation & Annotation

Standard YOLO dataset folder structure

data.yaml example:

train: ./dataset/images/train
val: ./dataset/images/val

nc: 2
names: ['person', 'car']

Recommended annotation tools: makesense.ai, labelImg, or Roboflow.

3. Training a Custom Model

from ultralytics import YOLO

model = YOLO("yolov8n.pt")  # or yolov8s.pt/m.pt

model.train(
    data="data.yaml",
    epochs=100,
    imgsz=640,
    batch=16,
    name="yolo26_custom"
)

4. Real-Time Inference Script

Full yolo_detect.py:

import cv2
import supervision as sv
from ultralytics import YOLO
import typer

app = typer.Typer(help="YOLO26 real-time detection")

model = YOLO("best.pt")  # Apna trained model path yahan dalo

def process(source: str, output_file: str, conf_threshold: float = 0.25):
    cap = cv2.VideoCapture(int(source) if source.isdigit() else source)
    if not cap.isOpened():
        raise typer.Exit("Source open nahi hua")

    width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
    height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
    fps = cap.get(cv2.CAP_PROP_FPS) or 30

    fourcc = cv2.VideoWriter_fourcc(*'mp4v')
    out = cv2.VideoWriter(output_file, fourcc, fps, (width, height))

    box_annotator = sv.BoundingBoxAnnotator()
    label_annotator = sv.LabelAnnotator()

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

        results = model(frame, conf=conf_threshold, verbose=False)[0]
        detections = sv.Detections.from_ultralytics(results)

        annotated = box_annotator.annotate(scene=frame.copy(), detections=detections)
        annotated = label_annotator.annotate(scene=annotated, detections=detections)

        out.write(annotated)
        cv2.imshow("YOLO26 Detection", annotated)

        if cv2.waitKey(1) & 0xFF == ord('q'):
            break

    cap.release()
    out.release()
    cv2.destroyAllWindows()

@app.command()
def run(source: str = "0", output: str = "output.mp4", conf: float = 0.25):
    process(source, output, conf)

if __name__ == "__main__":
    app()

5. Run Commands

# Webcam
python yolo_detect.py --source 0 --conf 0.3

# Video file
python yolo_detect.py --source video.mp4 --output result.mp4

Happy detecting! Don’t forget to like, share, and subscribe to the channel 🚀