Unknow/cours-ai-tutorials
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Initial commit

Browse Source
This commit is contained in:
Alexandre Gut
2026-03-31 13:28:59 +02:00
commit 7ec43ca17d
314 changed files with 189852 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
Divers/tutoriel25-3/README.md Normal file
View File

@@ -0,0 +1,8 @@
# Tutoriel 25
## Lecture des panneaux de limitation de vitesse
Les vidéos de ce tutoriel sont disponibles aux adresses suivantes:<br>
partie 1: https://www.youtube.com/watch?v=PvD5POjXw8Q <br>
partie 2: https://www.youtube.com/watch?v=TYDi0SNCUr0 <br>
partie 3: https://www.youtube.com/watch?v=fBysd-Y17Tw

67
Divers/tutoriel25-3/common.py Normal file
View File

@@ -0,0 +1,67 @@
import tensorflow as tf
from tensorflow.keras import layers, models
import os
import cv2
size=42
dir_images_panneaux="images_panneaux"
dir_images_autres_panneaux="images_autres_panneaux"
dir_images_sans_panneaux="images_sans_panneaux"
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

82
Divers/tutoriel25-3/dataset.py Normal file
View File

@@ -0,0 +1,82 @@
import numpy as np
import cv2
from multiprocessing import Pool
import multiprocessing
import random
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
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)
with Pool(nbr_thread) as p:
lot_result=p.map(modif_img, lot_original)
p.close()
return lot_result

40
Divers/tutoriel25-3/genere_fond.py Normal file
View File

@@ -0,0 +1,40 @@
import cv2
import numpy as np
import random
import os
import common
video="videos/France_Motorway.mp4"
if not os.path.isdir(common.dir_images_sans_panneaux):
os.mkdir(common.dir_images_sans_panneaux)
if not os.path.exists(video):
print("Vidéo non présente:", video)
quit()
cap=cv2.VideoCapture(video)
id=0
nbr_image=100000
nbr_image_par_frame=int(100000/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-common.size)
y=random.randint(0, h-common.size)
img=frame[y:y+common.size, x:x+common.size]
cv2.imwrite(common.dir_images_sans_panneaux+"/{:d}.png".format(id), img)
id+=1
if id==nbr_image:
quit()

29
Divers/tutoriel25-3/genere_panneaux.py Normal file
View File

@@ -0,0 +1,29 @@
import numpy as np
from sklearn.utils import shuffle
import cv2
import common
import dataset
tab_panneau, tab_image_panneau=common.lire_images_panneaux(common.dir_images_panneaux, common.size)
tab_images=np.array([]).reshape(0, common.size, common.size, 3)
tab_labels=[]
id=0
for image in tab_image_panneau:
lot=dataset.create_lot_img(image, 1000)
tab_images=np.concatenate([tab_images, lot])
tab_labels=np.concatenate([tab_labels, np.full(len(lot), id)])
id+=1
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])
tab_images, tab_labels=shuffle(tab_images, tab_labels)
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()

69
Divers/tutoriel25-3/lire_panneau.py Normal file
View File

@@ -0,0 +1,69 @@
import tensorflow as tf
import cv2
import os
import numpy as np
import random
import common
th1=30
th2=55
video_dir="d:/dashcam Cedric"
tab_panneau, tab_image_panneau=common.lire_images_panneaux(common.dir_images_panneaux)
model_panneau=common.panneau_model(len(tab_panneau))
checkpoint=tf.train.Checkpoint(model_panneau=model_panneau)
checkpoint.restore(tf.train.latest_checkpoint("./training_panneau/"))
l=os.listdir(video_dir)
random.shuffle(l)
for video in l:
if not video.endswith("mp4"):
continue
cap=cv2.VideoCapture(video_dir+"/"+video)
print("video:", video)
id_panneau=-1
while True:
ret, frame=cap.read()
if ret is False:
break
f_w, f_h, f_c=frame.shape
frame=cv2.resize(frame, (int(f_h/1.5), int(f_w/1.5)))
image=frame[200:400, 700:1000]
cv2.rectangle(frame, (700, 200), (1000, 400), (255, 255, 255), 1)
gray=cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
circles=cv2.HoughCircles(gray, cv2.HOUGH_GRADIENT, 1, 20, param1=th1, param2=th2, minRadius=5, maxRadius=45)
if circles is not None:
circles=np.int16(np.around(circles))
for i in circles[0,:]:
if i[2]!=0:
panneau=cv2.resize(image[max(0, i[1]-i[2]):i[1]+i[2], max(0, i[0]-i[2]):i[0]+i[2]], (common.size, common.size))/255
cv2.imshow("panneau", panneau)
prediction=model_panneau(np.array([panneau]), training=False)
print("Prediction:", prediction)
if np.any(np.greater(prediction[0], 0.6)):
id_panneau=np.argmax(prediction[0])
print(" -> C'est un panneau:", tab_panneau[id_panneau], "KM/H")
w, h, c=tab_image_panneau[id_panneau].shape
else:
print(" -> Ce n'est pas un panneau")
if id_panneau!=-1:
frame[0:h, 0:w, :]=tab_image_panneau[id_panneau]
cv2.putText(frame, "fichier:"+video, (30, 30), cv2.FONT_HERSHEY_DUPLEX, 1, (0, 255, 0), 1, cv2.LINE_AA)
cv2.imshow("Video", frame)
key=cv2.waitKey(1)&0xFF
if key==ord('q'):
quit()
if key==ord('a'):
for cpt in range(100):
ret, frame=cap.read()
if key==ord('f'):
break
cv2.destroyAllWindows()

140
Divers/tutoriel25-3/train_panneau.py Normal file
View File

@@ -0,0 +1,140 @@
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 common
import dataset
batch_size=128
nbr_entrainement=20
tab_panneau, tab_image_panneau=common.lire_images_panneaux(common.dir_images_panneaux, common.size)
tab_images=np.array([]).reshape(0, common.size, common.size, 3)
tab_labels=np.array([]).reshape(0, len(tab_image_panneau))
id=0
for image in tab_image_panneau:
lot=dataset.create_lot_img(image, 12000)
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(common.dir_images_autres_panneaux)
if files is None:
quit("Le repertoire d'image est vide:".format(common.dir_images_autres_panneaux))
nbr=0
for file in files:
if file.endswith("png"):
path=os.path.join(common.dir_images_autres_panneaux, file)
image=cv2.resize(cv2.imread(path), (common.size, common.size), cv2.INTER_LANCZOS4)
lot=dataset.create_lot_img(image, 700)
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=common.dir_images_sans_panneaux+"/{:d}.png".format(id)
if not os.path.isfile(file):
break
image=cv2.resize(cv2.imread(file), (common.size, common.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=common.panneau_model(len(tab_panneau))
checkpoint=tf.train.Checkpoint(model_panneau=model_panneau)
print("Entrainement")
train(train_ds, nbr_entrainement)
checkpoint.save(file_prefix="./training_panneau/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