[CNN] Custom Dataset으로 Pothole detection

캐글에서 데이터셋을 다운 받았다

Pothole Detection Dataset

다른 분들이 CNN으로 포트홀 탐지하는 코드를 올려주셔서 보면서 따라했다

 

 

필요한 라이브러리들 import !

import numpy as np
import pandas as pd
import tensorflow as tf
import cv2
import matplotlib.pyplot as plt
from keras.preprocessing.image import ImageDataGenerator
from keras.preprocessing import image

 

사진들 확인해보기

plt.imshow(cv2.imread("pothole_dataset/potholes/125.jpg"))

 

training set이랑 validation set으로 나누기 !

# data preprocessing

train_datagen = ImageDataGenerator(rescale = 1./255, shear_range = 0.2, zoom_range = 0.2, horizontal_flip = True, validation_split=0.2)
training_set = train_datagen.flow_from_directory('pothole_dataset', target_size = (64, 64), batch_size = 32,class_mode = 'binary',subset='training')
validation_set = train_datagen.flow_from_directory('pothole_dataset', target_size=(64, 64), batch_size=32, class_mode='binary', subset='validation')

 

CNN 빌딩 !

# bulding CNN

cnn = tf.keras.models.Sequential()
cnn.add(tf.keras.layers.Conv2D(filters=32, kernel_size=3, activation='relu', input_shape=[64, 64, 3])) # convolution
cnn.add(tf.keras.layers.MaxPool2D(pool_size=2, strides=2)) # pooling

## second conv
cnn.add(tf.keras.layers.Conv2D(filters=32, kernel_size=3, activation='relu'))
cnn.add(tf.keras.layers.MaxPool2D(pool_size=2, strides=2))

cnn.add(tf.keras.layers.Flatten()) # flattening

cnn.add(tf.keras.layers.Dense(units=128, activation='relu')) # full connection
cnn.add(tf.keras.layers.Dense(units=1, activation='sigmoid')) # output layer

 

CNN 학습시키기 !

# traing CNN

cnn.compile(optimizer = 'adam', loss = 'binary_crossentropy', metrics = ['accuracy']) # compiling
cnn.fit(x = training_set, validation_data = validation_set, epochs = 25) # training set -> training / validation set -> evaluating

 

test 결과 확인하기 !

test_img = 'pothole_dataset/potholes/184.jpg'  # another_images/normal/1.jpg 는 pothole으로 인식됨 (터널 입구 때문인 것 같음)
test_image = image.load_img(test_img, target_size = (64, 64))

plt.imshow(cv2.imread(test_img)) # test img 출력

test_image = image.img_to_array(test_image)
test_image = np.expand_dims(test_image, axis = 0)
result = cnn.predict(test_image)
training_set.class_indices

if result[0][0] == 1:
    prediction = 'pothole'
else:
    prediction = 'normal'
    
print(prediction)

 

 

---

 

일단 따라해보면서 코드를 이해해봤다

맨날 매트랩으로만 인공지능 머신러닝 해봤지 파이썬으로는 처음인데 사용하는 함수 몇몇개만 다르고 큰 틀은 똑같은 것 같다!!

 

팟팅 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!