树莓派4B+ YOLOv5+pytorch1.8+opencv4.3

qq_45613331

已于 2022-04-03 23:50:46 修改

阅读量2k

点赞数 1

文章标签： pytorch python opencv 目标检测

于 2022-02-20 14:20:24 首次发布

本文链接：https://blog.csdn.net/qq_45613331/article/details/123030530

版权

将Yolov5s部署树莓派4B上自己的过程和遇到的问题
先换源，再安装虚拟环境yolov5，参考树莓派4b安装python虚拟环境

再安装包，可以参考在树莓派部署Yolov5

树莓派终端输入 uname -a 确认自己的版本为armv71版本
树莓派终端输入 python3 确认自己的python版本为python3.7
执行安装命令

pip3 install https://wintics-opensource.s3.eu-west-3.amazonaws.com/torch-1.3.0a0%2Bdeadc27-cp37-cp37m-linux_armv7l.whl

树莓派终端输入 pip3 list 确认是否已经安装torch（大坑）

在这里插入图片描述

torch要>1.7，参考
树莓派4B部署yolov5：安装torch：torch1.8.0+torchvision0.9.0
安装成功

sudo会安装到本机，只有直接pip install才可以安装到虚拟环境。

tensorboard检测时可以不用安装，直接注释

opencv-python的安装（大坑）
去https://www.piwheels.org/simple/opencv-python/ 找到opencv_python-4.3.0.38-cp37-cp37m-linux_armv7l.whl下载

cd 到下载目录，download,

pip install opencv_python-4.3.0.38-cp37-cp37m-linux_armv7l.whl

安装成功，可以成功运行yolov5

cd .virtualenvs
source yolov5/bin/activate
cd /home/pi/Desktop/ZL-PI/yolov5-5.0
python detect.py
python detect.py --source 0

(yolov5) pi@raspberrypi:~/Desktop/ZL-PI/yolov5-5.0 $ python detect.py
Namespace(agnostic_nms=False, augment=False, classes=None, conf_thres=0.25, device='', exist_ok=False, img_size=640, iou_thres=0.45, name='exp', nosave=False, project='runs/detect', save_conf=False, save_txt=False, source='data/images', update=False, view_img=False, weights='yolov5s.pt')
YOLOv5 \U0001f680 2021-4-12 torch 1.8.0a0+56b43f4 CPU

Fusing layers... 
Model Summary: 224 layers, 7266973 parameters, 0 gradients
image 1/2 /home/pi/Desktop/ZL-PI/yolov5-5.0/data/images/bus.jpg: 640x480 4 persons, 1 bus, 1 fire hydrant, Done. (11.877s)
image 2/2 /home/pi/Desktop/ZL-PI/yolov5-5.0/data/images/zidane.jpg: 384x640 2 persons, 2 ties, Done. (9.523s)
Results saved to runs/detect/exp
Done. (22.209s)

8.camera
replace:
detect.py

 parser.add_argument('--source', type=str, default='0', help='source')

dataset:

if 'youtube.com/' in str(url) or 'youtu.be/' in str(url):

finish part1

pyqt5
important

https://blog.csdn.net/weixin_58006153/article/details/117745149
https://blog.csdn.net/idber/article/details/40076821

pip show sip

sudo -s
pip uninstall sip
rm -f /home/pi/.virtualenvs/yolov5/include/python3.7m/sip.h
rm -f /home/pi/.virtualenvs/yolov5/lib/python3.7/site-packages/sip*
rm -rf /home/pi/.virtualenvs/yolov5/include
exit


/home/pi/.local/share/Trash/files/PyQt5-5.15.6
/home/pi/.local/share/Trash/files/sip-4.19.25

sip

tar -zxvf sip-6.5.1.tar.gz
cd sip-6.5.1
python3 configure.py
sudo make
sudo make install

sip -V
>>>import sip
>>>print(sip.SIP_VERSION_STR)

cd PyQt-gpl-5.3.2
python3 configure.py

//https://pypi.org/project/PyQt5/5.15.5/#files
tar -zxvf sip-4.19.25.tar.gz
cd  sip-4.19.25
python3 configure.py
make
make install

//https://pypi.org/project/sip/4.19.6/#files
tar -zxvf PyQt5-5.15.5.tar.gz
cd PyQt5-5.15.5
python3 configure.py
make
make install
python
import PyQt5