app.py

import gradio as gr
import cv2
import os
import numpy as np
import pickle
from ultralytics import YOLO

# Charger le modèle entraîné
model = YOLO("My_best_model.pt")  # Adapter le chemin si besoin

# Dictionnaires de classes
class_names = {0: "suitcase", 1: "backpack", 2: "handbag"}
class_colors = {
    "suitcase": (255, 0, 0),
    "backpack": (0, 255, 0),
    "handbag": (0, 0, 255)
}

def process_video(input_video):
    # Charger la vidéo
    cap = cv2.VideoCapture(input_video)
    fps = int(cap.get(cv2.CAP_PROP_FPS))
    width, height = int(cap.get(3)), int(cap.get(4))

    # Fichiers temporaires
    output_path = "output_result.mp4"
    state_file = "state_temp.pkl"
    temp_txt = "final_counts.txt"
    out = cv2.VideoWriter(output_path, cv2.VideoWriter_fourcc(*'mp4v'), fps, (width, height))

    # État de suivi
    cumulative_counts = {cls: 0 for cls in class_names.values()}
    seen_ids = set()

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

        results = model.track(frame, persist=True, tracker="bytetrack.yaml")

        if results[0].boxes.id is not None:
            boxes = results[0].boxes.xyxy.cpu().numpy().astype(int)
            ids = results[0].boxes.id.cpu().numpy().astype(int)
            classes = results[0].boxes.cls.cpu().numpy().astype(int)
            confs = results[0].boxes.conf.cpu().numpy()

            for box, obj_id, cls, conf in zip(boxes, ids, classes, confs):
                cls_name = class_names.get(cls)
                if cls_name is None:
                    continue

                if obj_id not in seen_ids:
                    cumulative_counts[cls_name] += 1
                    seen_ids.add(obj_id)

                color = class_colors.get(cls_name, (0, 255, 0))
                x1, y1, x2, y2 = box
                label = f"{cls_name} ID:{obj_id} ({conf:.2f})"
                cv2.rectangle(frame, (x1, y1), (x2, y2), color, 2)
                cv2.putText(frame, label, (x1, y1 - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.6, color, 2)

        # Affichage du comptage
        y_offset = 30
        cv2.rectangle(frame, (10, 10), (300, 150), (0, 0, 0), -1)
        for cls_name, count in cumulative_counts.items():
            color = class_colors.get(cls_name, (0, 255, 0))
            cv2.putText(frame, f"{cls_name}: {count}", (20, y_offset), cv2.FONT_HERSHEY_SIMPLEX, 0.7, color, 2)
            y_offset += 30

        total = sum(cumulative_counts.values())
        cv2.putText(frame, f"Total: {total}", (20, y_offset + 10), cv2.FONT_HERSHEY_SIMPLEX, 0.9, (0, 255, 255), 2)

        out.write(frame)

    cap.release()
    out.release()

    # Sauvegarder les résultats dans un fichier texte
    with open(temp_txt, "w") as f:
        for k, v in cumulative_counts.items():
            f.write(f"{k}: {v}\n")
        f.write(f"Total: {total}\n")

    return output_path, temp_txt


# Interface Gradio
demo = gr.Interface(
    fn=process_video,
    inputs=gr.Video(label="Charger une vidéo"),
    outputs=[
        gr.Video(label="Vidéo annotée"),
        gr.File(label="Comptage final (fichier texte)")
    ],
    title="Détection et Comptage d'Objets avec YOLOv8",
    description="Téléversez une vidéo pour détecter et compter les valises, sacs à dos et sacs à main. Utilise YOLO + ByteTrack."
)

demo.launch()