tensorflow2.3实现卫星图像数据分类（CNN）

tensorflow2.3实现卫星图像数据分类

我们用CNN卷集神经网络实现卫星数据分类，数据有两种，lake 和 airplane
导入包

import tensorflow as tf
import numpy as np
import pathlib
import matplotlib.pyplot as plt

查看tensorflow的版本

print('Tensorflow version: {}'.format(tf.__version__))

Tensorflow version: 2.3

获取文件路径

data_dir = './2_class'

用pathlib构建路径对像

data_root = pathlib.Path(data_dir)

对目录进行迭代查看文件路径及对象

for item in data_root.iterdir():
    print(item)

/2_class/airplane
/2_class/lake

使用glob方法及正则表达式提取目录里面所有文件

all_image_path = list(data_root.glob('*/*'))

数据的数量

len(all_image_path)

1400

通过切片查看前2个文件

all_image_path[:2]

[PosixPath('/2_class/airplane/airplane_446.jpg'),
 PosixPath('/2_class/airplane/airplane_550.jpg')]

通过切片查看最后3个文件

all_image_path[-3:]

[PosixPath('/2_class/lake/lake_360.jpg'),
 PosixPath('/2_class/lake/lake_375.jpg'),
 PosixPath('/2_class/lake/lake_092.jpg')]

把文件目录转化为str形式

all_image_path = [str(path) for path in all_image_path]

显示其中的一些数据

all_image_path[10:12]

['/2_class/airplane/airplane_344.jpg',
 '/2_class/airplane/airplane_599.jpg']

由 PosixPath(’/2_class/lake/lake_360.jpg’)变成 ‘/2_class/airplane/airplane_599.jpg’

乱序图片

import random
random.shuffle(all_image_path)

图片数量

image_count = len(all_image_path)
image_count

提取目标值

label_names = sorted(item.name for item in data_root.glob('*/'))
label_names

['airplane', 'lake']

目标值airplane’和lake转换成训练时的目标值 0和1字典

label_to_index = dict((name,index) for index,name in enumerate(label_names))
label_to_index

{'airplane': 0, 'lake': 1}

pathlib.Path('/2_class/airplane/airplane_240.jpg').parent.name

'airplane'

把所有数据的标签转换

all_image_labels = [label_to_index[pathlib.Path(p).parent.name] for p in all_image_path]
all_image_labels[:5]

[0, 0, 0, 1, 1]

把标签0和1反转换为airplane’和lake

index_to_label = dict((v, k) for k, v in label_to_index.items())
index_to_label

{0: 'airplane', 1: 'lake'}

显示图片和label

import IPython.display as display
for n in range(3):             #随机显示3张图片
    image_index = random.choice(range(len(all_image_path)))
    display.display(display.Image(all_image_path[image_index]))
    print(index_to_label[all_image_labels[image_index]])
    print()

lake

lake

lake

加载单张图像

imag_path = all_image_paths[0]
imag_path

'dataset/2_class/lake/lake_700.jpg'

用 tf.io中的方法读取第一张图像

img_raw = tf.io.read_file(imag_path)
img_raw

<tf.Tensor: shape=(), dtype=string, numpy=b'\xff\xd8\xff\xe0\x00\x10JFIF\x00\x01\x01\x00\x00\x01\x00\x01\x00\x00\xff\xdb\x00C\x00\x08\x06\x06\x07\x06\x05\x08\x07\x07\x07\t\t\x08\n\x0c\x14\r\x0c\x0b\x0b\x0c\x19\x12\x13\x0f\x14\x1d\x1a\x1f\x1e\x1d\x1a\x1c\x1c $.\' ",#\x1c\x1c(7),01444\x1f\'9=82<.342\xff\xdb\x00C\x01\t\t\t\x0c\x0b\x0c\x18\r\r\x182!\x1c!22222222222222222222222222222222222222222222222222\xff\xc0\x00\x11\x08\x01\x00\x01\x00\x03\x01"\x00\x02\x11\x01\x03\x11\x01\xff\xc4\x00\x1f\x00\x00\x01\x05\x01\x01\x01\x01\x01\x01\x00\x00\x0

图像解码

img_tensor = tf.image.decode_image(img_raw)
img_tensor.shape

TensorShape([256, 256, 3])

图形数据类型

img_tensor.dtype

tf.uint8

自定义一个函数把前面几个方法一并完成，从读取到解码到大小转化再到类型转化最后归一化

def load_preprocess_image(img_path):
    img_raw = tf.io.read_file(img_path)
    img_tensor = tf.image.decode_jpeg(img_raw, channels=3)
    img_tensor = tf.image.resize(img_tensor, (256, 256))
    img_tensor = tf.cast(img_tensor, tf.float32)
    img = img_tensor/255
    return img

测试第500张

image_path = all_image_path[500]
plt.imshow(load_preprocess_image(image_path))

构建image图像的dataset

path_ds = tf.data.Dataset.from_tensor_slices(all_image_paths)
image_dataset = path_ds.map(load_preprocess_image)

构建label的dataset，label不需要map自定义的函数

label_dataset = tf.data.Dataset.from_tensor_slices(all_image_labels)

查看image_dataset的形状

image_dataset
label_dataset

<MapDataset shapes: (256, 256, 3), types: tf.float32>
<TensorSliceDataset shapes: (), types: tf.int32>

把image_dataset和label_dataset zip到一起（也不可以不用）

dataset = tf.data.Dataset.zip((image_dataset, label_dataset))
dataset

<ZipDataset shapes: ((256, 256, 3), ()), types: (tf.float32, tf.int32)>

设计训练数据和测试数据的数量

test_count = int(image_count*0.2)
train_count = image_count - test_count
test_count, train_count

(280, 1120)

创建训练集和测试集

train_dataset = dataset.skip(test_count)
test_dataset = dataset.take(test_count)

设置batch_size

BACH_SIZE = 32

设置训练数据输入的状况

train_dataset = train_dataset.repeat().shuffle(buffer_size=train_count).batch(BACH_SIZE)

设置测试数据测试的状况

test_dataset = test_dataset.batch(BACH_SIZE)

查看训练数据和测试数据的状况

test_dataset
train_dataset

<BatchDataset shapes: ((None, 256, 256, 3), (None,)), types: (tf.float32, tf.int32)>
<BatchDataset shapes: ((None, 256, 256, 3), (None,)), types: (tf.float32, tf.int32)>

建立模型

model = tf.keras.Sequential()
model.add(tf.keras.layers.Conv2D(64, (3, 3),input_shape=(256, 256, 3), activation='relu'))
model.add(tf.keras.layers.BatchNormalization())
model.add(tf.keras.layers.Conv2D(64, (3, 3), activation='relu'))
model.add(tf.keras.layers.BatchNormalization())
model.add(tf.keras.layers.Dropout(0.5))
model.add(tf.keras.layers.MaxPooling2D())
model.add(tf.keras.layers.Conv2D(128, (3, 3), activation='relu'))
model.add(tf.keras.layers.BatchNormalization())
model.add(tf.keras.layers.Dropout(0.5))
model.add(tf.keras.layers.Conv2D(128, (3, 3), activation='relu'))
model.add(tf.keras.layers.BatchNormalization())
model.add(tf.keras.layers.Dropout(0.5))
model.add(tf.keras.layers.MaxPooling2D())
model.add(tf.keras.layers.Conv2D(256, (3, 3), activation='relu'))
model.add(tf.keras.layers.BatchNormalization())
model.add(tf.keras.layers.Dropout(0.5))
model.add(tf.keras.layers.Conv2D(256, (3, 3), activation='relu'))
model.add(tf.keras.layers.BatchNormalization())
model.add(tf.keras.layers.Dropout(0.5))
model.add(tf.keras.layers.MaxPooling2D())
model.add(tf.keras.layers.Conv2D(512, (3, 3), activation='relu'))
model.add(tf.keras.layers.BatchNormalization())
model.add(tf.keras.layers.Dropout(0.5))
model.add(tf.keras.layers.MaxPooling2D())
model.add(tf.keras.layers.Conv2D(512, (3, 3), activation='relu'))
model.add(tf.keras.layers.BatchNormalization())
model.add(tf.keras.layers.Dropout(0.5))
model.add(tf.keras.layers.MaxPooling2D())
model.add(tf.keras.layers.Conv2D(1024, (3, 3), activation='relu'))
model.add(tf.keras.layers.BatchNormalization())
model.add(tf.keras.layers.Dropout(0.5))
model.add(tf.keras.layers.GlobalAveragePooling2D())
model.add(tf.keras.layers.Dense(1024, activation='relu'))
model.add(tf.keras.layers.BatchNormalization())
model.add(tf.keras.layers.Dense(256, activation='relu'))
model.add(tf.keras.layers.BatchNormalization())
model.add(tf.keras.layers.Dense(1, activation='sigmoid'))

模型概述

model.summary()

模型编译

model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['acc'])

设置批次

steps_per_epoch = train_count//BACH_SIZE
validation_step = test_count//BACH_SIZE

模型训练

history = model.fit(train_dataset, epochs=30, steps_per_epoch=steps_per_epoch, validation_data=test_dataset, validation_steps=validation_step)

模型训练结果

plt.plot(history.epoch, history.history.get('loss'), label='loss')
plt.plot(history.epoch, history.history.get('val_loss'), label='val_loss')
plt.legend()

plt.plot(history.epoch, history.history.get('acc'), label='loss')
plt.plot(history.epoch, history.history.get('val_acc'), label='val_acc')
plt.legend()

参考文献：
https://study.163.com/course/introduction/1004573006.htm