一、构建自己的数据集
1、格式必须为jpg、jpeg或png。
2、在models/research/object_detection文件夹下创建images文件夹，在images文件夹下创建train和val两个文件夹，分别存放训练集图片和测试集图片。
3、下载labelImg目标检测标注工具
（1）下载地址：https://github.com/tzutalin/labelImg
（2）由于LabelImg是用Python编写的，并使用Qt作为其图形界面。
因此，python2安装qt4:
sudo apt-get install pyqt4-dev-tools
python3安装qt5：
sudo apt-get install pyqt5-dev-tools
（3）安装lxml
sudo apt-get install python-lxml
（4）解压，进入LabelImg-master文件夹，使用make编译
cd labelImg-master
make all
（5）打开LabelImg
python labelImg.py
（6）使用LabelImg
- 使用Ctrl + u分别加载models/research/object_detection/images中train和val两个文件夹里的图像。
- 使用Ctrl + r选择xml文件保存的地址，对应地选择保存在train和val文件夹即可。
- 使用w创建一个矩形框，标注完一张图片中的所有物体后，Ctrl + s保存即可生成该图片对应的xml文件。
4、创建xml_to_csv.py并运行
分别将train和val文件夹下的xml文件生成对应的csv文件，并将csv文件拷贝到models/research/object_detection/data中。
xml_to_csv.py如下，以train为例。
import os
import glob
import pandas as pd
import xml.etree.ElementTree as ET
pathStr = r'/home/somnus/boat/train'
os.chdir(pathStr)
def xml_to_csv(path):
xml_list = []
for xml_file in glob.glob(path + '/*.xml'):
tree = ET.parse(xml_file)
root = tree.getroot()
for member in root.findall('object'):
value = (root.find('filename').text,
int(root.find('size').find('width').text),
int(root.find('size').find('height').text),
member.find('name').text,
int(member.find('bndbox').find('xmin').text),
int(member.find('bndbox').find('ymin').text),
int(member.find('bndbox').find('xmax').text),
int(member.find('bndbox').find('ymax').text)
)
xml_list.append(value)
column_name = ['filename', 'width', 'height', 'class', 'xmin', 'ymin', 'xmax', 'ymax']
xml_df = pd.DataFrame(xml_list, columns=column_name)
return xml_df
def main():
#image_path = os.path.join(os.getcwd(), 'annotations')
image_path = pathStr
xml_df = xml_to_csv(image_path)
xml_df.to_csv('boat_train.csv', index=None)
print('Successfully converted xml to csv.')
main()
5、创建generate_tfrecord.py并运行，以train为例，从而生成对应的TFrecord数据文件。
"""
Usage:
# From tensorflow/models/
# Create train data:
python generate_tfrecord.py --csv_input=data/train_labels.csv --output_path=train.record # Create test data:
python generate_tfrecord.py --csv_input=data/test_labels.csv --output_path=test.record
"""
from __future__ import division
from __future__ import print_function
from __future__ import absolute_import import os
import io
import pandas as pd
import tensorflow as tf from PIL import Image
from object_detection.utils import dataset_util
from collections import namedtuple, OrderedDict #########根据需要修改路径
os.chdir('/home/somnus/models/research/object_detection') flags = tf.app.flags
flags.DEFINE_string('csv_input', '', 'Path to the CSV input')
flags.DEFINE_string('output_path', '', 'Path to output TFRecord')
flags.DEFINE_string('image_dir', '', 'Path to images')
FLAGS = flags.FLAGS ####根据需要修改标签
def class_text_to_int(row_label):
if row_label == 'car':
return 1
else:
None def split(df, group):
data = namedtuple('data', ['filename', 'object'])
gb = df.groupby(group)
return [data(filename, gb.get_group(x)) for filename, x in zip(gb.groups.keys(), gb.groups)] def create_tf_example(group, path):
with tf.gfile.GFile(os.path.join(path, '{}'.format(group.filename)), 'rb') as fid:
encoded_jpg = fid.read()
encoded_jpg_io = io.BytesIO(encoded_jpg)
image = Image.open(encoded_jpg_io)
width, height = image.size filename = group.filename.encode('utf8')
image_format = b'jpg'
xmins = []
xmaxs = []
ymins = []
ymaxs = []
classes_text = []
classes = [] for index, row in group.object.iterrows():
xmins.append(row['xmin'] / width)
xmaxs.append(row['xmax'] / width)
ymins.append(row['ymin'] / height)
ymaxs.append(row['ymax'] / height)
classes_text.append(row['class'].encode('utf8'))
classes.append(class_text_to_int(row['class'])) tf_example = tf.train.Example(features=tf.train.Features(feature={
'image/height': dataset_util.int64_feature(height),
'image/width': dataset_util.int64_feature(width),
'image/filename': dataset_util.bytes_feature(filename),
'image/source_id': dataset_util.bytes_feature(filename),
'image/encoded': dataset_util.bytes_feature(encoded_jpg),
'image/format': dataset_util.bytes_feature(image_format),
'image/object/bbox/xmin': dataset_util.float_list_feature(xmins),
'image/object/bbox/xmax': dataset_util.float_list_feature(xmaxs),
'image/object/bbox/ymin': dataset_util.float_list_feature(ymins),
'image/object/bbox/ymax': dataset_util.float_list_feature(ymaxs),
'image/object/class/text': dataset_util.bytes_list_feature(classes_text),
'image/object/class/label': dataset_util.int64_list_feature(classes),
}))
return tf_example def main(_):
writer = tf.python_io.TFRecordWriter(FLAGS.output_path) #####根据需要修改训练集或测试集图片路径
path = os.path.join('images/train') examples = pd.read_csv(FLAGS.csv_input)
grouped = split(examples, 'filename')
for group in grouped:
tf_example = create_tf_example(group, path)
writer.write(tf_example.SerializeToString()) writer.close()
output_path = os.path.join(os.getcwd(), FLAGS.output_path)
print('Successfully created the TFRecords: {}'.format(output_path)) if __name__ == '__main__':
tf.app.run()
运行generate_tfrecord.py
python generate_tfrecord.py --csv_input=data/car_train.csv --output_path=data/car_train.record
python generate_tfrecord.py --csv_input=data/car_val.csv --output_path=data/car_val.record
二、准备配置文件
1、在models/research/object_detection/data文件夹下创建mymodel_label_map.pbtxt文件，可以模仿pet_label_map.pbtxt，内容修改为自己模型识别的标签，从1开始编号。
item {
id: 1
name: 'car'
}
2、在object_detection下创建training文件夹，在models/research/object_detection/samples/configs中找到需要的模型文件，并拷贝到training文件夹下，以ssd_mobilenet_v1_coco.config为例。
model {
ssd {
#根据需要修改训练的数据类数
num_classes: 1
box_coder {
faster_rcnn_box_coder {
y_scale: 10.0
x_scale: 10.0
height_scale: 5.0
width_scale: 5.0
}
}
matcher {
argmax_matcher {
matched_threshold: 0.5
unmatched_threshold: 0.5
ignore_thresholds: false
negatives_lower_than_unmatched: true
force_match_for_each_row: true
}
}
similarity_calculator {
iou_similarity {
}
}
anchor_generator {
ssd_anchor_generator {
num_layers: 6
min_scale: 0.2
max_scale: 0.95
aspect_ratios: 1.0
aspect_ratios: 2.0
aspect_ratios: 0.5
aspect_ratios: 3.0
aspect_ratios: 0.3333
}
}
image_resizer {
fixed_shape_resizer {
height: 300
width: 300
}
}
box_predictor {
convolutional_box_predictor {
min_depth: 0
max_depth: 0
num_layers_before_predictor: 0
use_dropout: false
dropout_keep_probability: 0.8
kernel_size: 1
box_code_size: 4
apply_sigmoid_to_scores: false
conv_hyperparams {
activation: RELU_6,
regularizer {
l2_regularizer {
weight: 0.00004
}
}
initializer {
truncated_normal_initializer {
stddev: 0.03
mean: 0.0
}
}
batch_norm {
train: true,
scale: true,
center: true,
decay: 0.9997,
epsilon: 0.001,
}
}
}
}
feature_extractor {
type: 'ssd_mobilenet_v1'
min_depth: 16
depth_multiplier: 1.0
conv_hyperparams {
activation: RELU_6,
regularizer {
l2_regularizer {
weight: 0.00004
}
}
initializer {
truncated_normal_initializer {
stddev: 0.03
mean: 0.0
}
}
batch_norm {
train: true,
scale: true,
center: true,
decay: 0.9997,
epsilon: 0.001,
}
}
}
loss {
classification_loss {
weighted_sigmoid {
}
}
localization_loss {
weighted_smooth_l1 {
}
}
hard_example_miner {
num_hard_examples: 3000
iou_threshold: 0.99
loss_type: CLASSIFICATION
max_negatives_per_positive: 3
min_negatives_per_image: 0
}
classification_weight: 1.0
localization_weight: 1.0
}
normalize_loss_by_num_matches: true
post_processing {
batch_non_max_suppression {
score_threshold: 1e-8
iou_threshold: 0.6
max_detections_per_class: 100
max_total_detections: 100
}
score_converter: SIGMOID
}
}
}
train_config: {
#根据需要修改训练批次
batch_size: 24
optimizer {
rms_prop_optimizer: {
learning_rate: {
exponential_decay_learning_rate {
initial_learning_rate: 0.004
decay_steps: 800720
decay_factor: 0.95
}
}
momentum_optimizer_value: 0.9
decay: 0.9
epsilon: 1.0
}
}
#这两行注释
#fine_tune_checkpoint: "PATH_TO_BE_CONFIGURED/model.ckpt"
#from_detection_checkpoint: true
# Note: The below line limits the training process to 200K steps, which we
# empirically found to be sufficient enough to train the pets dataset. This
# effectively bypasses the learning rate schedule (the learning rate will
# never decay). Remove the below line to train indefinitely.
num_steps: 200000
data_augmentation_options {
random_horizontal_flip {
}
}
data_augmentation_options {
ssd_random_crop {
}
}
}
train_input_reader: {
tf_record_input_reader {
#修改路径
input_path: "data/car_train.record"
}
#修改路径
label_map_path: "data/mymodel_label_map.pbtxt"
}
eval_config: {
num_examples: 200
# Note: The below line limits the evaluation process to 10 evaluations.
# Remove the below line to evaluate indefinitely.
max_evals: 10
}
eval_input_reader: {
tf_record_input_reader {
#修改路径
input_path: "data/car_val.record"
}
#修改路径
label_map_path: "data/mymodel_label_map.pbtxt"
shuffle: false
num_readers: 1
}
3、在models/research/object_detection下运行
python ./legacy/train.py --logtostderr --train_dir=training/ --pipeline_config_path=training/ssd_mobilenet_v1_coco.config
三、生成可被调用的模型
python export_inference_graph.py --input_type image_tensor --pipeline_config_path training/ssd_mobilenet_v1_coco.config --trained_checkpoint_prefix training/model.ckpt-8004 --output_directory car_inference_graph
其中，model.ckpt-后面的数字可以看training文件夹下的文件，选个最大的数字；--output_directory=指定的是模型生成的文件夹名字，可根据需要修改。
参考