jetson nano 部署yoloV3,yoloV4,yoloV3-tiny,yoloV4-tiny

转载自：jetson nano 部署yoloV3,yoloV4,yoloV3-tiny,yoloV4-tiny_dingding的专栏-CSDN博客

jetson nano 部署yoloV3,yoloV4,yoloV3-tiny,yoloV4-tiny
VIP文章 Miss yang 2020 2020-12-20 18:47:08 1405 收藏 9
分类专栏：深度学习
版权

系统：ubuntu 自带cuda10.0

1、下载与安装darknet

git clone https://github.com/AlexeyAB/darknet
cd darknet

2、以下步骤我都在直接进入commit中操作

（1）核对cuda，若输入nvcc -V 出现命令位找到，则说明未配置cuda，继续阅读，否则直接执行第3步

nvcc -V
-bash: nvcc: command not found

（2）将cuda加入路径

sudo chmod -R 777 ~/.bashrc
gedit ~/.bashrc

（3）在文件最后加入以下三行

    export CUBA_HOME=/usr/local/cuda-10.2
    export LD_LIBRARY_PATH=/usr/local/cuda-10.2/lib64:$LD_LIBRARY_PATH
    export PATH=/usr/local/cuda-10.2/bin:$PATH

（4）执行使cuda生效

source ~/.bashrc

（5）可以输入nvcc -V查看版本

3、修改Makefile文件

cd darknet
gedit Makefile

进入Makefile，修改

    GPU=1
    CUDNN=1
    OPENCV=１ # 如果安装了opencv则改为１
    OPENMP=0
    DEBUG=0
    ...
    NVCC=/usr/local/cuda-10.0/bin/nvcc # 原为 NVCC = nvcc，cuda后面的紧跟自己的cuda版本

4、输入make -j2

5、在darknet文件夹中查看是否有yolov4-tiny.weight文件（其他版本类似），cfg里是否有coco.data，yolov4-tiny.cfg(其他版本类似),

6、接下来可以显示了

    ./darknet detect cfg/yolov4-tiny.cfg yolov4-tiny.weights data/dog.jpg # 简写版

    ./darknet detector test cfg/coco.data cfg/yolov4-tiny.cfg yolov4-tiny.weights data/dog.jpg # 完整版

视频检测

./darknet detector demo cfg/coco.data cfg/yolov4-tiny.cfg yolov4-tiny.weights data/test_video.mp4

测试 Yolov4-tiny CSI摄像头实时检测

    # 使用CSI摄像头前，要安装GStreamer
    sudo add-apt-repository universe
    sudo add-apt-repository multiverse
    sudo apt-get update
    sudo apt-get install gstreamer1.0-tools gstreamer1.0-alsa gstreamer1.0-plugins-base gstreamer1.0-plugins-good gstreamer1.0-plugins-bad gstreamer1.0-plugins-ugly gstreamer1.0-libav
    sudo apt-get install libgstreamer1.0-dev libgstreamer-plugins-base1.0-dev libgstreamer-plugins-good1.0-dev libgstreamer-plugins-bad1.0-dev
    # 配置GStreamer管道，并使用Yolov4-tiny模型进行检测
    ./darknet detector demo cfg/coco.data cfg/yolov4-tiny.cfg yolov4-tiny.weights "nvarguscamerasrc ! video/x-raw(memory:NVMM), width=1280, height=720, format=NV12, framerate=30/1 ! nvvidconv flip-method=0 ! video/x-raw, width=1280, height=720, format=BGRx ! videoconvert ! video/x-raw, format=BGR ! appsink"

测试usb摄像头（未测）

darknet detector demo cfg/coco.data cfg/yolov4.cfg yolov4.weights

jetson nano 部署yoloV3,yoloV4,yoloV3-tiny,yoloV4-tiny

猜你喜欢