数据集制作:
重点:把数据变成需要的数据格式record形式文件
1.搜集图片
本人采用的是kinect2相机搜集的225张图片,其中170为训练集,50张为测试集,5为验证集
2.安装标注labelIm.exe
https://github.com/tzutalin/labelImg
具体软件用法可参开百度说明
3.打标签
![](https://img.laitimes.com/img/_0nNw4CM6IyYiwiM6ICdiwiIwczX0xiRGZkRGZ0Xy9GbvNGL2EzXlpXazxSP9E1T1UlaNFzYU9ke4wmYwhGWhxGZzwEMW1mY1RzRapnTtxkb5ckYplTeMZTTINGMShUYfRHelRHLwEzX39GZhh2css2RkBnVHFmb1clWvB3MaVnRtp1XlBXe0xyayFWbyVGdhd3LcV2Zh1Wa9M3clN2byBXLzN3btg3Pn5GcukDOzQzMxcTM5ETMwkTMwIzLc52YucWbp5GZzNmLn9Gbi1yZtl2Lc9CX6MHc0RHaiojIsJye.png)
最终形成如下:
170张这样的文件,每个文件记录假的图片位置和标准框的大小,以74.xml为例
<annotation>
<folder>pictures_data</folder>
<filename>74.png</filename>
<path>C:\Users\Administrator\Desktop\pictures_data\74.png</path>
<source>
<database>Unknown</database>
</source>
<size>
<width>960</width>
<height>540</height>
<depth>3</depth>
</size>
<segmented>0</segmented>
<object>
<name>book</name>
<pose>Unspecified</pose>
<truncated>0</truncated>
<difficult>0</difficult>
<bndbox>
<xmin>358</xmin>
<ymin>185</ymin>
<xmax>490</xmax>
<ymax>353</ymax>
</bndbox>
</object>
<object>
<name>paper</name>
<pose>Unspecified</pose>
<truncated>0</truncated>
<difficult>0</difficult>
<bndbox>
<xmin>412</xmin>
<ymin>362</ymin>
<xmax>527</xmax>
<ymax>488</ymax>
</bndbox>
</object>
<object>
<name>coffee</name>
<pose>Unspecified</pose>
<truncated>0</truncated>
<difficult>0</difficult>
<bndbox>
<xmin>570</xmin>
<ymin>271</ymin>
<xmax>674</xmax>
<ymax>420</ymax>
</bndbox>
</object>
</annotation>
4.形成训练的数据文件
- xml_to_csv.py 把文件改成一个csv文件
- python generate_tfrecord.py --csv_input=images\train_labels.csv --image_dir=images\train --output_path=train.record python generate_tfrecord.py --csv_input=images\test_labels.csv --image_dir=images\test --output_path=test.record