下载
代码和MobileNet训练模型可以从以下位置下载:
https://github.com/djmv/MobilNet_SSD_opencv
https://github.com/chuanqi305/MobileNet-SSD
http://www.ebenezertechs.com/mobilenet-ssd-using-opencv-3-4-1-deep-learning-module-python/
https://github.com/djmv/MobilNet_SSD_opencv
网友加速
在Raspberry Pi上设置TensorFlow对象检测API的教程
https://github.com/EdjeElectronics/TensorFlow-Object-Detection-on-the-Raspberry-Pi#6-detect-objects
https://github.com/EdjeElectronics/TensorFlow-Object-Detection-on-the-Raspberry-Pi
https://www.youtube.com/watch?v=gGqVNuYol6o&feature=youtu.be
https://blog.csdn.net/weixin_43558453/article/details/85175253
Tensorflow官方提供的本地编译的方式在arm嵌入式设备运行Tensorflow Lite
https://blog.csdn.net/weixin_43558453/article/details/86507764
即使是在实时检测并亮灯的时候树莓派的CPU的占用率也65%左右,所以小小的树莓派用Tengine还是有可以继续发掘的潜力的。
如果大家对Tengine框架的性能有兴趣可以参考一下我之前写的那篇文章,关于Tengine和
树莓派实现目标实时检测opencv-Moblenet
Tengine 推断引擎:树莓派也能玩转深度学习
http://shumeipai.nxez.com/2018/12/07/tengine-inference-engine-raspberry-pi-deep-learning.html
可以看到单帧耗时有所下降(400ms-700ms),
使用opencvd的级联器
https://www.cs.cmu.edu/~efros/courses/LBMV07/Papers/viola-cvpr-01.pdf