使用tensorflow object detection API 训练自己的目标检测模型

參考文章 : 

https://blog.csdn.net/zj1131190425/article/details/80711857

https://blog.csdn.net/zj1131190425/article/details/80726353

https://blog.csdn.net/zj1131190425/article/details/80778888

==================

問題排除 

如果出现  ImportError: No module named 'object_detection' 这种问题

 在   F:\Anaconda\Lib\site-packages  这个文件夹目录下，写一个.pth文件，内容为两条路径

C:\Users\Ricky\Desktop\models-master\research
C:\Users\Ricky\Desktop\models-master\research\slim

(https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/detection_model_zoo.md)

下载的是faster_rcnn_inception_v2_coco模型

搭配 C:\Users\Ricky\Desktop\models-master\research\object_detection\samples\configs 裡面的

faster_rcnn_inception_v2_pets

protoc ./object_detection/protos/*.proto --python_out=.    # 如果不行，前面添加路徑

C:\tensorflow-master\tensorflow\models>protoc object_detection/protos/*.proto --python_out=.