diff --git a/server-trainer/Camera_Identification_VitesseRoadSign-ModelTraining.py b/server-trainer/Camera_Identification_VitesseRoadSign-ModelTraining.py
new file mode 100644
index 0000000..ee8a039
--- /dev/null
+++ b/server-trainer/Camera_Identification_VitesseRoadSign-ModelTraining.py
@@ -0,0 +1,208 @@
+import tensorflow as tf
+from tensorflow.keras import layers, models
+import numpy as np
+from sklearn.utils import shuffle
+from sklearn.model_selection import train_test_split
+import cv2
+import os
+import time
+import dataset
+
+size=42
+dir_images_panneaux="server-trainer/images/road_sign_speed_trainers/panneaux"
+dir_images_autres_panneaux="server-trainer/images/road_sign_speed_trainers/autres_panneaux"
+dir_images_genere_sans_panneaux="server-trainer/images/road_sign_speed_trainers/genere_sans_panneaux"
+
+batch_size=128
+nbr_entrainement=1 #20
+
+def panneau_model(nbr_classes):
+    model=tf.keras.Sequential()
+
+    model.add(layers.Input(shape=(size, size, 3), dtype='float32'))
+    
+    model.add(layers.Conv2D(128, 3, strides=1))
+    model.add(layers.Dropout(0.2))
+    model.add(layers.BatchNormalization())
+    model.add(layers.Activation('relu'))
+
+    model.add(layers.Conv2D(128, 3, strides=1))
+    model.add(layers.Dropout(0.2))
+    model.add(layers.BatchNormalization())
+    model.add(layers.Activation('relu'))
+
+    model.add(layers.MaxPool2D(pool_size=2, strides=2))
+
+    model.add(layers.Conv2D(256, 3, strides=1))
+    model.add(layers.Dropout(0.3))
+    model.add(layers.BatchNormalization())
+    model.add(layers.Activation('relu'))
+
+    model.add(layers.Conv2D(256, 3, strides=1))
+    model.add(layers.Dropout(0.4))
+    model.add(layers.BatchNormalization())
+    model.add(layers.Activation('relu'))
+
+    model.add(layers.MaxPool2D(pool_size=2, strides=2))
+
+    model.add(layers.Flatten())
+    model.add(layers.Dense(512, activation='relu'))
+    model.add(layers.Dropout(0.5))
+    model.add(layers.BatchNormalization())
+    model.add(layers.Dense(nbr_classes, activation='sigmoid'))
+    
+    return model
+
+def lire_images_panneaux(dir_images_panneaux, size=None):
+    tab_panneau=[]
+    tab_image_panneau=[]
+
+    if not os.path.exists(dir_images_panneaux):
+        quit("Le repertoire d'image n'existe pas: {}".format(dir_images_panneaux))
+
+    files=os.listdir(dir_images_panneaux)
+    if files is None:
+        quit("Le repertoire d'image est vide: {}".format(dir_images_panneaux))
+
+    for file in sorted(files):
+        if file.endswith("png"):
+            tab_panneau.append(file.split(".")[0])
+            image=cv2.imread(dir_images_panneaux+"/"+file)
+            if size is not None:
+                image=cv2.resize(image, (size, size), cv2.INTER_LANCZOS4)
+            tab_image_panneau.append(image)
+            
+    return tab_panneau, tab_image_panneau
+
+tab_panneau, tab_image_panneau=lire_images_panneaux(dir_images_panneaux, size)
+
+tab_images=np.array([]).reshape(0, size, size, 3)
+tab_labels=np.array([]).reshape(0, len(tab_image_panneau))
+
+id=0
+for image in tab_image_panneau:
+    lot = []
+    for _ in range(120):
+        lot.append(dataset.modif_img(image))
+    lot = np.array(lot)
+    tab_images=np.concatenate((tab_images, lot))
+    tab_labels=np.concatenate([tab_labels, np.repeat([np.eye(len(tab_image_panneau))[id]], len(lot), axis=0)])
+    id += 1
+
+
+files=os.listdir(dir_images_autres_panneaux)
+if files is None:
+    quit("Le repertoire d'image est vide:".format(dir_images_autres_panneaux))
+
+nbr=0
+for file in files:
+    lot = []
+    if file.endswith("png"):
+        path=os.path.join(dir_images_autres_panneaux, file)
+        image=cv2.resize(cv2.imread(path), (size, size), cv2.INTER_LANCZOS4)
+        for _ in range(700):
+            lot.append(dataset.modif_img(image))
+        lot = np.array(lot)
+        tab_images=np.concatenate([tab_images, lot])
+        nbr+=len(lot)
+
+tab_labels=np.concatenate([tab_labels, np.repeat([np.full(len(tab_image_panneau), 0)], nbr, axis=0)])
+
+nbr_np=int(len(tab_images)/2)
+
+id=1
+nbr=0
+tab=[]
+for cpt in range(nbr_np):
+    file=dir_images_genere_sans_panneaux+"/{:d}.png".format(id)
+    if not os.path.isfile(file):
+        break
+    image=cv2.resize(cv2.imread(file), (size, size))
+    tab.append(image)
+    id+=1
+    nbr+=1
+
+tab_images=np.concatenate([tab_images, tab])
+tab_labels=np.concatenate([tab_labels, np.repeat([np.full(len(tab_image_panneau), 0)], nbr, axis=0)])
+
+tab_panneau=np.array(tab_panneau)
+tab_images=np.array(tab_images, dtype=np.float32)/255
+tab_labels=np.array(tab_labels, dtype=np.float32) #.reshape([-1, 1])
+
+train_images, test_images, train_labels, test_labels=train_test_split(tab_images, tab_labels, test_size=0.10)
+
+train_ds=tf.data.Dataset.from_tensor_slices((train_images, train_labels)).batch(batch_size)
+test_ds=tf.data.Dataset.from_tensor_slices((test_images, test_labels)).batch(batch_size)
+
+print("train_images", len(train_images))
+print("test_images", len(test_images))
+
+@tf.function
+def train_step(images, labels):
+  with tf.GradientTape() as tape:
+    predictions=model_panneau(images)
+    loss=my_loss(labels, predictions)
+  gradients=tape.gradient(loss, model_panneau.trainable_variables)
+  optimizer.apply_gradients(zip(gradients, model_panneau.trainable_variables))
+  train_loss(loss)
+  train_accuracy(labels, predictions)
+
+def train(train_ds, nbr_entrainement):
+  for entrainement in range(nbr_entrainement):
+    start=time.time()
+    for images, labels in train_ds:
+      train_step(images, labels)
+    message='Entrainement {:04d}: loss: {:6.4f}, accuracy: {:7.4f}%, temps: {:7.4f}'
+    print(message.format(entrainement+1,
+                         train_loss.result(),
+                         train_accuracy.result()*100,
+                         time.time()-start))
+    train_loss.reset_states()
+    train_accuracy.reset_states()
+    test(test_ds)
+
+def my_loss(labels, preds):
+    labels_reshape=tf.reshape(labels, (-1, 1))
+    preds_reshape=tf.reshape(preds, (-1, 1))
+    result=loss_object(labels_reshape, preds_reshape)
+    return result
+    
+def test(test_ds):
+  start=time.time()
+  for test_images, test_labels in test_ds:
+    predictions=model_panneau(test_images)
+    t_loss=my_loss(test_labels, predictions)
+    test_loss(t_loss)
+    test_accuracy(test_labels, predictions)
+  message='   >>> Test: loss: {:6.4f}, accuracy: {:7.4f}%, temps: {:7.4f}'
+  print(message.format(test_loss.result(),
+                       test_accuracy.result()*100,
+                       time.time()-start))
+  test_loss.reset_states()
+  test_accuracy.reset_states()
+
+optimizer=tf.keras.optimizers.Adam()
+loss_object=tf.keras.losses.BinaryCrossentropy()
+train_loss=tf.keras.metrics.Mean()
+train_accuracy=tf.keras.metrics.BinaryAccuracy()
+test_loss=tf.keras.metrics.Mean()
+test_accuracy=tf.keras.metrics.BinaryAccuracy()
+model_panneau=panneau_model(len(tab_panneau))
+checkpoint=tf.train.Checkpoint(model_panneau=model_panneau)
+
+print("Entrainement")
+train(train_ds, nbr_entrainement)
+checkpoint.save(file_prefix="server-ia/data/modeles/road_sign_speed_trainers/modele_panneau")
+
+quit()
+
+for i in range(len(test_images)):
+    prediction=model_panneau(np.array([test_images[i]]))
+    print("prediction", prediction[0])
+    if np.sum(prediction[0])<0.6:
+        print("Ce n'est pas un panneau")
+    else:
+        print("C'est un panneau:", tab_panneau[np.argmax(prediction[0])])
+    cv2.imshow("image", test_images[i])
+    if cv2.waitKey()&0xFF==ord('q'):
+        break
diff --git a/server-trainer/RoadSigns-ModelTraining_MobileNetV2.py b/server-trainer/RoadSigns-ModelTraining_MobileNetV2.py
new file mode 100644
index 0000000..dceda31
--- /dev/null
+++ b/server-trainer/RoadSigns-ModelTraining_MobileNetV2.py
@@ -0,0 +1,99 @@
+import os
+import tensorflow as tf
+from tensorflow.keras.preprocessing.image import ImageDataGenerator
+from tensorflow.keras.applications import MobileNetV2
+from tensorflow.keras.layers import AveragePooling2D, Dropout, Flatten, Dense, Input
+from tensorflow.keras.models import Model
+from tensorflow.keras.optimizers import Adam
+from sklearn.model_selection import train_test_split
+from sklearn.metrics import classification_report
+from imutils import paths
+import numpy as np
+
+# Définir les chemins des données d'entraînement et de validation
+train_data_dir = "server-trainer/images/road_sign_trainers/train_speed"
+valid_data_dir = "server-trainer/images/road_sign_trainers/test_speed"
+
+# Initialiser les hyperparamètres
+INIT_LR = 1e-4
+EPOCHS = 1
+BS = 32
+
+# Charger et prétraiter les images
+train_datagen = ImageDataGenerator(
+    rescale=1.0 / 255,
+    rotation_range=20,
+    zoom_range=0.15,
+    width_shift_range=0.2,
+    height_shift_range=0.2,
+    shear_range=0.15,
+    horizontal_flip=True,
+    fill_mode="nearest"
+)
+
+valid_datagen = ImageDataGenerator(rescale=1.0 / 255)
+
+train_generator = train_datagen.flow_from_directory(
+    train_data_dir,
+    target_size=(224, 224),
+    batch_size=BS,
+    class_mode='categorical'
+)
+
+valid_generator = valid_datagen.flow_from_directory(
+    valid_data_dir,
+    target_size=(224, 224),
+    batch_size=BS,
+    class_mode='categorical'
+)
+
+# Générer le fichier class_names.txt
+class_names_file = os.path.join("server-ia/data/modeles/RoadSign", "class_names.txt")
+with open(class_names_file, "w") as f:
+    # Utiliser le générateur class_indices pour récupérer les noms de classe et les écrire dans le fichier
+    for class_name, class_index in train_generator.class_indices.items():
+        f.write(f"{class_name}\n")
+
+# Charger le modèle pré-entraîné MobileNetV2 sans la couche supérieure
+base_model = MobileNetV2(
+    weights="imagenet",
+    include_top=False,
+    input_tensor=Input(shape=(224, 224, 3))
+)
+
+# Construire le modèle de tête qui sera placé au-dessus du modèle de base
+head_model = base_model.output
+head_model = AveragePooling2D(pool_size=(7, 7))(head_model)
+head_model = Flatten(name="flatten")(head_model)
+head_model = Dense(128, activation="relu")(head_model)
+head_model = Dropout(0.5)(head_model)
+head_model = Dense(len(train_generator.class_indices), activation="softmax")(head_model)
+
+# Combiner le modèle de base avec le modèle de tête
+model = Model(inputs=base_model.input, outputs=head_model)
+
+# Geler les couches du modèle de base
+for layer in base_model.layers:
+    layer.trainable = False
+
+# Compiler le modèle
+opt = Adam(learning_rate=INIT_LR)
+model.compile(loss="categorical_crossentropy", optimizer=opt, metrics=["accuracy"])
+
+# Entraîner le modèle
+history = model.fit(
+    train_generator,
+    steps_per_epoch=len(train_generator),
+    validation_data=valid_generator,
+    validation_steps=len(valid_generator),
+    epochs=EPOCHS
+)
+
+# Sauvegarder le modèle
+model.save("server-ia/data/modeles/RoadSign/modele_signaux_routiers.h5")
+
+# Évaluer le modèle
+print("[INFO] Évaluation du modèle...")
+predictions = model.predict(valid_generator, steps=len(valid_generator), verbose=1)
+predictions = np.argmax(predictions, axis=1)
+print(classification_report(valid_generator.classes, predictions, target_names=valid_generator.class_indices.keys()))
diff --git a/server-trainer/RoadSigns-ModelTraining_Yolovv3.py b/server-trainer/RoadSigns-ModelTraining_Yolovv3.py
new file mode 100644
index 0000000..27c3ef9
--- /dev/null
+++ b/server-trainer/RoadSigns-ModelTraining_Yolovv3.py
@@ -0,0 +1,31 @@
+import os
+
+# Assurez-vous d'être dans le répertoire Darknet
+os.chdir("chemin/vers/votre/darknet")
+
+# Entraînement du modèle YOLOv3 pour les panneaux de signalisation routière
+# Utilisation du CPU uniquement
+
+# Chemin vers les données d'entraînement et de validation
+train_data = "server-trainer/images/road_sign_trainers/train_speed"
+valid_data = "server-trainer/images/road_sign_trainers/test_speed"
+
+# Paramètres d'entraînement
+batch_size = 64
+subdivisions = 16
+num_epochs = 1000
+
+# Commande d'entraînement
+train_command = f"./darknet detector train data/obj.data cfg/yolov3_custom_train.cfg yolov3.weights -map -dont_show -gpus 0"
+
+# Boucle d'entraînement
+for epoch in range(num_epochs):
+    print(f"Epoch {epoch+1}/{num_epochs}")
+    
+    # Entraînement sur les données d'entraînement
+    os.system(train_command)
+
+    # Validation sur les données de validation
+    os.system(f"./darknet detector map data/obj.data cfg/yolov3_custom_test.cfg backup/yolov3_custom_train_{epoch+1}.weights")
+
+print("Entraînement terminé!")
diff --git a/server-trainer/UsersIdentification-ModelTraining.py b/server-trainer/UsersIdentification-ModelTraining.py
new file mode 100644
index 0000000..37eb7d0
--- /dev/null
+++ b/server-trainer/UsersIdentification-ModelTraining.py
@@ -0,0 +1,33 @@
+import cv2
+import os
+import numpy as np
+import pickle
+
+image_dir="server-trainer/images/avatars/"
+current_id=0
+label_ids={}
+x_train=[]
+y_labels=[]
+
+for root, dirs, files in os.walk(image_dir):
+    if len(files):
+        label=root.split("/")[-1]
+        for file in files:
+            if file.endswith("jpg"):
+                path=os.path.join(root, file)
+                if not label in label_ids:
+                    label_ids[label]=current_id
+                    current_id+=1
+                id_=label_ids[label]
+                image=cv2.imread(path, cv2.IMREAD_GRAYSCALE)
+                x_train.append(image)
+                y_labels.append(id_)
+
+with open("server-ia/data/modeles/camera_identification_user/labels.pickle", "wb") as f:
+    pickle.dump(label_ids, f)
+
+x_train=np.array(x_train)
+y_labels=np.array(y_labels)
+recognizer=cv2.face.LBPHFaceRecognizer_create()
+recognizer.train(x_train, y_labels)
+recognizer.save("server-ia/data/modeles/camera_identification_user/trainner.yml")
diff --git a/server-trainer/dataset.py b/server-trainer/dataset.py
new file mode 100644
index 0000000..d83c49e
--- /dev/null
+++ b/server-trainer/dataset.py
@@ -0,0 +1,82 @@
+import numpy as np
+import cv2
+import multiprocessing
+import random
+# dataset pour train_panneau.py
+def bruit(image_orig):
+    h, w, c = image_orig.shape
+    n = np.random.randn(h, w, c) * random.randint(5, 30)
+    return np.clip(image_orig + n, 0, 255).astype(np.uint8)
+
+def change_gamma(image, alpha=1.0, beta=0.0):
+    return np.clip(alpha * image + beta, 0, 255).astype(np.uint8)
+    
+def modif_img(img):
+    h, w, c = img.shape
+
+    r_color = [np.random.randint(255), np.random.randint(255), np.random.randint(255)]
+    img = np.where(img == [142, 142, 142], r_color, img).astype(np.uint8)
+
+    if np.random.randint(3):
+        k_max = 3
+        kernel_blur = np.random.randint(k_max) * 2 + 1
+        img = cv2.GaussianBlur(img, (kernel_blur, kernel_blur), 0)
+
+    M = cv2.getRotationMatrix2D((int(w / 2), int(h / 2)), random.randint(-10, 10), 1)
+    img = cv2.warpAffine(img, M, (w, h))
+        
+    if np.random.randint(2):
+        a = int(max(w, h) / 5) + 1
+        pts1 = np.float32([[0, 0], [w, 0], [0, h], [w, h]])
+        pts2 = np.float32([[0 + random.randint(-a, a), 0 + random.randint(-a, a)], [w - random.randint(-a, a), 0 + random.randint(-a, a)], [0 + random.randint(-a, a), h - random.randint(-a, a)], [w - random.randint(-a, a), h - random.randint(-a, a)]])        
+        M = cv2.getPerspectiveTransform(pts1,pts2)
+        img = cv2.warpPerspective(img, M, (w, h))
+        
+    if np.random.randint(2):
+        r = random.randint(0, 5)
+        h2 = int(h * 0.9)
+        w2 = int(w * 0.9)
+        if r == 0:
+            img = img[0:w2, 0:h2]
+        elif r == 1:
+            img = img[w - w2:w, 0:h2]
+        elif r == 2:
+            img = img[0:w2, h - h2:h]
+        elif r == 3:
+            img = img[w - w2:w, h - h2:h]
+        img = cv2.resize(img, (h, w))
+
+    if np.random.randint(2):
+        r = random.randint(1, int(max(w, h) * 0.15))
+        img = img[r:w - r, r:h - r]
+        img = cv2.resize(img, (h, w))
+
+    if not np.random.randint(4):
+        t = np.empty((h, w, c), dtype=np.float32)
+        for i in range(h):
+            for j in range(w):
+                for k in range(c):
+                    t[i][j][k] = (i / h)
+        M = cv2.getRotationMatrix2D((int(w / 2), int(h / 2)), np.random.randint(4) * 90, 1)
+        t = cv2.warpAffine(t, M, (w, h))
+        img = (cv2.multiply((img / 255).astype(np.float32), t) * 255).astype(np.uint8)
+
+    img = change_gamma(img, random.uniform(0.6, 1.0), -np.random.randint(50))
+
+    if not np.random.randint(4):
+        p = (15 + np.random.randint(10)) / 100
+        img = (img * p + 50 * (1 - p)).astype(np.uint8) + np.random.randint(100)
+
+    img = bruit(img)
+        
+    return img
+
+# NOT WORKING ERREUR
+def create_lot_img(image, nbr, nbr_thread=None):
+    if nbr_thread is None:
+        nbr_thread = multiprocessing.cpu_count()
+    lot_original = np.repeat([image], nbr, axis=0)
+    p = Pool(nbr_thread)
+    lot_result = p.map(modif_img, lot_original)
+    p.close()
+    return lot_result
\ No newline at end of file
diff --git a/server-trainer/genere_images_panneaux.py b/server-trainer/genere_images_panneaux.py
new file mode 100644
index 0000000..2c17c7d
--- /dev/null
+++ b/server-trainer/genere_images_panneaux.py
@@ -0,0 +1,85 @@
+import tensorflow as tf
+from tensorflow.keras import layers, models
+import numpy as np
+from sklearn.utils import shuffle
+import os
+import cv2
+import dataset
+###GENERE DES IMAGES AVEC PANNEAUX DEPUIS UN MEDIA###
+# Tuto25[OpenCV] Lecture des panneaux de vitesse p.1 (houghcircles) 28min
+# Taille des images
+size = 42
+# Chemin vers le répertoire contenant les images de panneaux
+dir_images_panneaux = "server-trainer/images/road_sign_speed_trainers/panneaux"
+dir_images_genere_panneaux = "server-trainer/images/road_sign_speed_trainers/genere_panneaux"
+
+# Fonction pour lire les images de panneaux à partir du répertoire spécifié
+def lire_images_panneaux(dir_images_panneaux, size=None):
+    tab_panneau = []
+    tab_image_panneau = []
+
+    # Vérifie si le répertoire existe
+    if not os.path.exists(dir_images_panneaux):
+        quit("Le repertoire d'image n'existe pas: {}".format(dir_images_panneaux))
+
+    # Liste tous les fichiers dans le répertoire
+    files = os.listdir(dir_images_panneaux)
+    
+    # Quitte si le répertoire est vide
+    if files is None:
+        quit("Le repertoire d'image est vide: {}".format(dir_images_panneaux))
+
+    # Parcours des fichiers dans le répertoire
+    for file in sorted(files):
+        # Vérifie si le fichier est un fichier PNG
+        if file.endswith("png"):
+            # Ajoute le nom du fichier (sans l'extension) à tab_panneau
+            tab_panneau.append(file.split(".")[0])
+            
+            # Lit l'image et redimensionne si une taille est spécifiée
+            image = cv2.imread(dir_images_panneaux + "/" + file)
+            if size is not None:
+                image = cv2.resize(image, (size, size), cv2.INTER_LANCZOS4)
+            tab_image_panneau.append(image)
+            
+    return tab_panneau, tab_image_panneau
+
+# Appel de la fonction pour lire les images de panneaux
+tab_panneau, tab_image_panneau = lire_images_panneaux(dir_images_panneaux, size)
+
+# Initialisation des tableaux pour les images et les labels
+tab_images = np.array([]).reshape(0, size, size, 3)
+tab_labels = []
+
+# Génération des données d'entraînement
+id = 0
+for image in tab_image_panneau:
+    lot = []
+    for _ in range(100):
+        lot.append(dataset.modif_img(image))
+    lot = np.array(lot)
+    tab_images = np.concatenate([tab_images, lot])
+    tab_labels = np.concatenate([tab_labels, np.full(len(lot), id)])
+    id += 1
+
+# Conversion des tableaux en type approprié et normalisation des images
+tab_panneau = np.array(tab_panneau)
+tab_images = np.array(tab_images, dtype=np.float32) / 255
+tab_labels = np.array(tab_labels).reshape([-1, 1])
+
+# Mélange des données
+tab_images, tab_labels = shuffle(tab_images, tab_labels)
+
+# Boucle pour sauvegarder les images générées dans le répertoire de sortie
+for i in range(len(tab_images)):
+    # Générer un nom de fichier unique en fonction de l'ID et du nom du panneau
+    file_name = "{:d}_{}.png".format(i, tab_panneau[int(tab_labels[i])])
+    # Enregistrer l'image dans le répertoire de sortie
+    cv2.imwrite(os.path.join(dir_images_genere_panneaux, file_name), tab_images[i] * 255.0)  # Retour à l'échelle 0-255
+
+# Affichage des images avec leur label
+for i in range(len(tab_images)):
+    cv2.imshow("panneau", tab_images[i])
+    print("label", tab_labels[i], "panneau", tab_panneau[int(tab_labels[i])])
+    if cv2.waitKey() & 0xFF == ord('q'):
+        quit()
\ No newline at end of file
diff --git a/server-trainer/genere_images_sans_panneaux.py b/server-trainer/genere_images_sans_panneaux.py
new file mode 100644
index 0000000..c7d76c6
--- /dev/null
+++ b/server-trainer/genere_images_sans_panneaux.py
@@ -0,0 +1,39 @@
+import cv2
+import numpy as np
+import random
+import os
+###GENERE DES IMAGES SANS PANNEAUX DEPUIS UN MEDIA###
+#YT:Tuto25[Tensorflow2] Lecture des panneaux de vitesse p.2 - 4min30
+
+size=42
+video="server-trainer/videos/autoroute.mp4"
+dir_images_genere_sans_panneaux="server-trainer/images/road_sign_speed_trainers/genere_sans_panneaux"
+
+if not os.path.isdir(dir_images_genere_sans_panneaux):
+    os.mkdir(dir_images_genere_sans_panneaux)
+
+if not os.path.exists(video):
+    print("Vidéo non présente:", video)
+    quit()
+    
+cap=cv2.VideoCapture(video)
+
+id=0
+nbr_image=1500
+
+nbr_image_par_frame=int(1500/cap.get(cv2.CAP_PROP_FRAME_COUNT))+1
+
+while True:
+    ret, frame=cap.read()
+    if ret is False:
+        quit()
+    h, w, c=frame.shape
+
+    for cpt in range(nbr_image_par_frame):
+        x=random.randint(0, w-size)
+        y=random.randint(0, h-size)
+        img=frame[y:y+size, x:x+size]        
+        cv2.imwrite(dir_images_genere_sans_panneaux+"/{:d}.png".format(id), img)
+        id+=1
+        if id==nbr_image:
+            quit()
\ No newline at end of file