目标检测实战(一)--使用TensorFlow Object Detection API检测自己的图像数据集
下载并使用labelimg
网上有很多关于labelimg的下载方法
labelimg下载链接
打开labelimg效果如下
左边的工具栏说明:
1.open是打开单张图片
2.open dir是打开一个包含图像的文件夹
3.change save dir是选定你图片标注好之后,标注文件保存的位置
4.next image和prev image分别代表下张图片和上张图片
5.save是每此一张图片标注完之后保存
6.PascalVOC或者yolo可以来回切换保存格式,一般用前者(为xml)
7.Creat/nrectBox表示画矩形框
下载安装TensorFlow Object Detection API
处理数据
图片是从网上采集的,通过labelimg我们得到了多个xml标注文件文件(每张图片的),我们必须把它们转换成tfrecord文件。
首先使用以下代码(xml_to_csv.py):
# -*- coding: utf-8 -*- """ Created on Thu Oct 10 14:20:54 2019 @author: asus """ import os import glob import pandas as pd import xml.etree.ElementTree as ET 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')[0].text), int(root.find('size')[1].text), member[0].text, int(member[4][0].text), int(member[4][1].text), int(member[4][2].text), int(member[4][3].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(): os.chdir(r"C:\Users\asus\mouse_project_file\annotations") image_path = os.path.join(os.getcwd(), 'train_xml') xml_df = xml_to_csv(image_path) os.chdir(r"C:\Users\asus\mouse_project_file\data") xml_df.to_csv('train_labels.csv', index=None) print('Successfully converted xml to csv.') main()
注意:
需要将main函数里面的输入文件和输出文件改成自己的;
xml_to_csv函数的参数为输入文件(就是xml所在文件夹);
最后输出的是一个csv文件(我的csv文件名叫train_labels用于训练,同时也可以生成test_labels用于测试),如下图:
进入该代码文件位置,执行python xml_to_csv.py
无异常会输出:Successfully converted xml to csv.
接下来将csv文件转换为tfrecord文件,执行以下代码
""" Created on Thu Oct 10 14:34:30 2019 @author: asus """ """ 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 """ # -*- coding: utf-8 -*- 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 flags = tf.app.flags flags.DEFINE_string('csv_input', '', 'Path to the CSV input') flags.DEFINE_string('output_path', '', 'Path to output TFRecord') FLAGS = flags.FLAGS # TO-DO replace this with label map def class_text_to_int(row_label): if row_label == 'belt': return 1 elif row_label == 'milk': return 2 else: return 3 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(_): os.chdir(r"C:\Users\asus\mouse_project_file\data") writer = tf.python_io.TFRecordWriter(FLAGS.output_path) os.chdir(r"C:\Users\asus\mouse_project_file") path = os.path.join(os.getcwd(), r'images\test_images') print(path) 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() os.chdir(r"C:\Users\asus\mouse_project_file\data") output_path = os.path.join(os.getcwd(), FLAGS.output_path) print('Successfully created the TFRecords: {}'.format(output_path)) if __name__ == '__main__': tf.app.run()
注意修改class_text_to_int函数,将自己的类别修改到编号
进入到相应的目录,执行
#Create train data:
python generate_tfrecord.py --csv_input=train_labels.csv --output_path=train.record
#Create test data:
python generate_tfrecord.py --csv_input=test_labels.csv --output_path=test.record
这里的csv_input为csv文件,output_path为输出的tfrecord文件,这样就得到了train.record文件和test.record文件
参数配置
这里首先配置一个pbtxt文件(建议下载一个Notepad,查看代码,文件比较方便),具体内容如下:
如果有多个类别,就写多个item就行了
接下来配置模型参数文件
不同的模型有不同的参数配置文件,这里我们使ssd_mobilenet_v1_coco
模型下载链接:
模型文件model zoo
这个链接里面有许多模型可供下载
同时在tensorflow\models\research\object_detection\samples\configs有每个模型对应的config文件
模型下载解压之后,得到ssd_mobilenet_v1_coco_2018_01_28模型文件夹,然后找到对应的config文件tensorflow\models\research\object_detection\samples\configs\ssd_mobilenet_v1_coco.config
config中有几个必须要修改的参数:
1.num_classes:设置为你的类别数量
2.batch_size:设置batch值
3.fine_tune_checkpoint:预训练模型位置
4.num_steps:总步数
5.input_path:record训练或测试文件位置
6.label_map_path:pbtxt文件位置
具体参数配置内容如下:
# SSD with Mobilenet v1 configuration for MSCOCO Dataset. # Users should configure the fine_tune_checkpoint field in the train config as # well as the label_map_path and input_path fields in the train_input_reader and # eval_input_reader. Search for "PATH_TO_BE_CONFIGURED" to find the fields that # should be configured. model { ssd { 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: { batch_size: 6 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 } } fine_tune_checkpoint: "mouse_project_file/model/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: 9000 data_augmentation_options { random_horizontal_flip { } } data_augmentation_options { ssd_random_crop { } } } train_input_reader: { tf_record_input_reader { input_path: "mouse_project_file/data/train.record" } label_map_path: "mouse_project_file/training/label.pbtxt" } eval_config: { num_examples: 30 # Note: The below line limits the evaluation process to 10 evaluations. # Remove the below line to evaluate indefinitely. max_evals: 20 } eval_input_reader: { tf_record_input_reader { input_path: "mouse_project_file/data/test.record" } label_map_path: "mouse_project_file/training/label.pbtxt" shuffle: False num_readers: 1 }
可以在以上的基础上修改
开始训练
进入(cd)tensorflow\models\research\object_detection\legacy有一个train.py文件
执行以下语句:
python object_detection/legacy/train.py \ --logtostderr \ --train_dir=mouse_project_file\result \ --pipeline_config_path=mouse_project_file/ssd_mobilenet_v1_coco.config
train_dir为模型保存路径
pipeline_config_path为参数配置路径
其它参数的具体含义可以查看train.py文件
测试及可视化
在训练过程中,你也可以测试模型的效果,打开另一个窗口,执行以下语句:
python eval.py \ --logtostderr \ --pipeline_config_path=mouse_project_file/ssd_mobilenet_v1_coco.config \ --checkpoint_dir=mouse_project_file\result \ --eval_dir=mouse_project_file/eval
pipeline_config_path:参数配置文件
checkpoint_dir:模型保存文件夹
eval_dir:测试结果保存文件
其它参数配置,详见eval.py文件
同时,你也可以进行可视化,这是对测试结果的可视化,执行以下命令:
tensorboard \ --logdir=mouse_project_file/eval
也可以对训练结果可视化:
tensorboard \ --logdir=mouse_project_file/result
上图是测试集可视化的效果
导出模型
python object_detection/export_inference_graph.py --input_type=image_tensor --pipeline_config_path=mouse_project_file\ssd_mobilenet_v1_fpn_shared_box_predictor_640x640_coco14_sync.config --trained_checkpoint_prefix=mouse_project_file\ssd_mobilenet_v1_fpn_shared_box_predictor_640x640_coco14_sync_result\model.ckpt-2996 --output_directory=mouse_project_file\frz_out
1.input_type:输入节点类型,一共有三种类型(
image_tensor,
encoded_image_string_tensor,
tf_example)
2.pipeline_config_path:模型对应的配置文件路径
3.trained_checkpoint_prefix:模型文件位置
4.output_directory:导出文件位置
其它参数详见export_inference_graph.py
导出样例如下:
训练效果
参考文献:
1.https://blog.csdn.net/qq_38593211/article/details/82823255
2.https://blog.csdn.net/Kalenee/article/details/80629262
- 点赞
- 收藏
- 分享
- 文章举报
- 使用tensorflow object detection API 训练自己的目标检测模型 (三)
- Linux(Ubuntu18.4、GPU服务器)Tensorflow object detection API使用配置教程(特别详细)tensorflow目标检测教程
- 从0.1开始Tensorflow object detection API使用教程(特别详细)tensorflow目标检测教程
- 使用Tensorflow Object Detection API识别自己的图像
- 使用TensorFlow Object Detection API进行图像物体检测
- 目标检测(Google object_detection) API 上训练自己的数据集
- Tensorflow object_detection API 目标检测环境搭建
- tensorflow object detection api训练自己的数据集
- TensorFlow Object Detection API教程——制作自己的数据集
- (更新视频教程)Tensorflow object detection API 搭建属于自己的物体识别模型(2)——训练并使用自己的模型
- 基于win10,GPU的Tensorflow Object Detection API部署及USB摄像头目标检测
- 目标检测 - Tensorflow Object Detection API
- tensorflow object_detection 用自己的数据训练目标检测模型Mobilenet
- (三论文)目标检测 - Tensorflow Object Detection API几种模型的对比
- Windows下安装TensorFlow Object Detection API,训练自己的数据集
- 使用谷歌Object Detection API进行目标检测、训练新的模型(使用VOC 2012数据集) (转)
- TensorFlow Object Detection API教程——利用自己制作的数据集进行训练预测和测试
- Windows下安装TensorFlow Object Detection API,训练自己的数据集
- 利用TensorFlow Object Detection API的mask-rcnn模型 训练自己的样本数据集
- 计算机视觉实时目标检测 TensorFlow Object Detection API