首页 » 机器视觉 » 正文
机器视觉

darknet 训练流程

1、darknet编译

  1. git clone https://github.com/pjreddie/darknet.git
  2. cd darknet
  3. 修改Makefile
GPU=1  #加GPU
CUDNN=1  #
OPENCV=1  #加opencv
OPENMP=0 
DEBUG=1

4、make

有时候make时候会报错，下面列举一些遇到的报错信息

1、

./src/convolutional_kernels.cu -o obj/convolutional_kernels.o
/bin/sh: 1: nvcc: not found
Makefile:89: recipe for target 'obj/convolutional_kernels.o' failed
make: *** [obj/convolutional_kernels.o] Error 12

上面报错信息喝nvcc有关，修改Makefile如下：
NVCC
NVCC为自己安装cuda时候的位置。
2、

./src/image_opencv.cpp:12:1: error: ‘IplImage’ 报错

以前安装的是opencv3系列的没有出现上述问题，当安装opencv4系列的就会遇到上述问题，主要是因为在opencv安装时候3系列和4系列会有些不同。
opencv4安装：
加入依赖：

sudo apt-get install build-essential
sudo apt-get install cmake git libgtk2.0-dev pkg-config libavcodec-dev libavformat-dev libswscale-dev

cmake -D CMAKE_BUILD_TYPE=Release -D CMAKE_INSTALL_PREFIX=/usr/local/myself_path -D OPENCV_GENERATE_PKGCONFIG=ON ..

重点：OPENCV_GENERATE_PKGCONFIG=ON，这样在/usr/local/myself_path/lib能生成opencv4.pc文件。

sudo cp /usr/local/lib/pkgconfig/opencv4.pc /usr/lib/pkgconfig

sudo mv /usr/lib/pkgconfig/opencv4.pc /usr/lib/pkgconfig/opencv.pc 

make -j8
sudo make install

将opencv4.pc 改成opencv.pc，否则make时候会找不到opencv
重点：配置opencv的环境变量

sudo vim /etc/ld.so.conf.d/opencv.conf

输入/usr/local/myself_path

sudo vim /etc/profile

在末尾加入：

export PKG_CONFIG_PATH=/usr/local/lib/pkgconfig:$PKG_CONFIG_PATH
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/myself_path/lib

更新环境变量：sudo source /etc/profile

sudo vim /etc/bash.bashrc

export PKG_CONFIG_PATH=$PKG_CONFIG_PATH:/usr/local/lib/pkgconfig
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/myself_path/lib

更新环境变量：sudo source ~/.bashrc
更新系统缓存：sudo ldconfig
opencv安装完成后修改darknet源码：

cd darknet/src

vim image_opencv.cpp

附上完整代码，替换就好

#ifdef OPENCV

#include "stdio.h"
#include "stdlib.h"
#include "opencv2/opencv.hpp"
#include "image.h"

using namespace cv;

extern "C" {

Mat image_to_mat(image im)
{
    image copy = copy_image(im);
    constrain_image(copy);
    if(im.c == 3) rgbgr_image(copy);

    Mat m(cv::Size(im.w,im.h), CV_8UC(im.c));
    int x,y,c;

    int step = m.step;
    for(y = 0; y < im.h; ++y){
        for(x = 0; x < im.w; ++x){
            for(c= 0; c < im.c; ++c){
                float val = im.data[c*im.h*im.w + y*im.w + x];
                m.data[y*step + x*im.c + c] = (unsigned char)(val*255);
            }
        }
    }

    free_image(copy);
    return m;
}

image mat_to_image(Mat m)
{
    int h = m.rows;
    int w = m.cols;
    int c = m.channels();
    image im = make_image(w, h, c);
    unsigned char *data = (unsigned char *)m.data;
    int step = m.step;
    int i, j, k;

    for(i = 0; i < h; ++i){
        for(k= 0; k < c; ++k){
            for(j = 0; j < w; ++j){
                im.data[k*w*h + i*w + j] = data[i*step + j*c + k]/255.;
            }
        }
    }
    rgbgr_image(im);
    return im;
}

void *open_video_stream(const char *f, int c, int w, int h, int fps)
{
    VideoCapture *cap;
    if(f) cap = new VideoCapture(f);
    else cap = new VideoCapture(c);
    if(!cap->isOpened()) return 0;
    if(w) cap->set(CAP_PROP_FRAME_WIDTH, w);
    if(h) cap->set(CAP_PROP_FRAME_HEIGHT, w);
    if(fps) cap->set(CAP_PROP_FPS, w);
    return (void *) cap;
}

image get_image_from_stream(void *p)
{
    VideoCapture *cap = (VideoCapture *)p;
    Mat m;
    *cap >> m;
    if(m.empty()) return make_empty_image(0,0,0);
    return mat_to_image(m);
}

image load_image_cv(char *filename, int channels)
{
    int flag = -1;
    if (channels == 0) flag = -1;
    else if (channels == 1) flag = 0;
    else if (channels == 3) flag = 1;
    else {
        fprintf(stderr, "OpenCV can't force load with %d channels\n", channels);
    }
    Mat m;
    m = imread(filename, flag);
    if(!m.data){
        fprintf(stderr, "Cannot load image \"%s\"\n", filename);
        char buff[256];
        sprintf(buff, "echo %s >> bad.list", filename);
        system(buff);
        return make_image(10,10,3);
        //exit(0);
    }
    image im = mat_to_image(m);
    return im;
}

int show_image_cv(image im, const char* name, int ms)
{
    Mat m = image_to_mat(im);
    imshow(name, m);
    int c = waitKey(ms);
    if (c != -1) c = c%256;
    return c;
}

void make_window(char *name, int w, int h, int fullscreen)
{
    namedWindow(name, WINDOW_NORMAL); 
    if (fullscreen) {
        setWindowProperty(name, WND_PROP_FULLSCREEN, WINDOW_FULLSCREEN);
    } else {
        resizeWindow(name, w, h);
        if(strcmp(name, "Demo") == 0) moveWindow(name, 0, 0);
    }
}

}

#endif

通过上述方法来解决./src/image_opencv.cpp:12:1: error: ‘IplImage’ 报错问题

2、darknet训练自己数据集

在darknet文件夹下创建自己的数据集myData

/darknet/$mkdir myData

在myData文件夹下创建三个文件夹：
在这里插入图片描述
其中annotations下存放xml文件(自己标注的格式是xml格式)
JPEGImages下存放标注的图像
在ImageSets下创建Main文件夹存放图像的名字(.txt)
脚本代码：

import os

src_path = "/darknet/myData/annotations/"
img_path = os.listdir(src_path)
save_path = "/darknet/myData/ImageSets/Main/train.txt"
for img_path1 in img_path:
    if img_path1[-4:] == ".xml":
        img_path_txt = img_path1[:-4]
        with open(save_path,"a+") as f:
            f.write(img_path_txt + "\n")

将xml格式转换为darknet需要的格式，利用darknet/scripts文件夹下的voc_label.py，代码需要更改一下

import xml.etree.ElementTree as ET
import pickle
import os
from os import listdir, getcwd
from os.path import join
sets=[('myData', 'train'), ('myData', 'val')]  #替换为自己的数据集
classes = ["class1", "class2", "class3"]     #修改为自己的类别
 
def convert(size, box):
    dw = 1./(size[0])
    dh = 1./(size[1])
    x = (box[0] + box[1])/2.0 - 1
    y = (box[2] + box[3])/2.0 - 1
    w = box[1] - box[0]
    h = box[3] - box[2]
    x = x*dw
    w = w*dw
    y = y*dh
    h = h*dh
    return (x,y,w,h)
def convert_annotation(year, image_id):
    in_file = open('myData/annotations/%s.xml'%(image_id))  #将数据集放于当前目录下
    out_file = open('myData/labels/%s.txt'%(image_id), 'w')
    tree=ET.parse(in_file)
    root = tree.getroot()
    size = root.find('size')
    w = int(size.find('width').text)
    h = int(size.find('height').text)
    for obj in root.iter('object'):
        difficult = obj.find('difficult').text
        cls = obj.find('name').text
        if cls not in classes or int(difficult)==1:
            continue
        cls_id = classes.index(cls)
        xmlbox = obj.find('bndbox')
        b = (float(xmlbox.find('xmin').text), float(xmlbox.find('xmax').text), float(xmlbox.find('ymin').text), float(xmlbox.find('ymax').text))
        bb = convert((w,h), b)
        out_file.write(str(cls_id) + " " + " ".join([str(a) for a in bb]) + '\n')
wd = getcwd()
for year, image_set in sets:
    if not os.path.exists('myData/labels/'):
        os.makedirs('myData/labels/')
    image_ids = open('myData/ImageSets/Main/%s.txt'%(image_set)).read().strip().split()
    list_file = open('myData/%s_%s.txt'%(year, image_set), 'w')
    for image_id in image_ids:
        list_file.write('%s/myData/JPEGImages/%s.jpg\n'%(wd,image_id))
        convert_annotation(year, image_id)
    list_file.close()   
#os.system("cat 2007_train.txt 2007_val.txt > train.txt")     #修改为自己的数据集用作训练

添加voc.data，目前将voc.data和voc.names 存放在mydata下

classes= 3    #修改为自己的类别数
train = /darknet/myData/myData_train.txt
valid = /darknet/myData/myData_val.txt
names = /darknet/myData/voc.names
backup = /darknet/backup

添加voc.names

class1
class2
class3

在/darknet下下载预权重

/darknet/$ wget https://pjreddie.com/media/files/darknet53.conv.74

修改cfg/yolov3-voc.cfg

[net]
# Testing
 batch=64
 subdivisions=32   #每批训练的个数=batch/subvisions，根据自己GPU显存进行修改，显存不够改大一些
# Training
# batch=64
# subdivisions=16
width=416
height=416
channels=3
momentum=0.9
decay=0.0005
angle=0
saturation = 1.5
exposure = 1.5
hue=.1
 
learning_rate=0.001
burn_in=1000
max_batches = 50200  #训练步数
policy=steps
steps=40000,45000  #开始衰减的步数
scales=.1,.1
 
 
 
[convolutional]
batch_normalize=1
filters=32
size=3
stride=1
pad=1
activation=leaky
 
.....
 
[convolutional]
size=1
stride=1
pad=1
filters=24   #filters = 3 * ( classes + 5 )   here,filters=3*(3+5)
activation=linear
 
[yolo]
mask = 6,7,8
anchors = 10,13,  16,30,  33,23,  30,61,  62,45,  59,119,  116,90,  156,198,  373,326
classes=3    #修改为自己的类别数
num=9
jitter=.3
ignore_thresh = .5
truth_thresh = 1
random=1
 
[route]
layers = -4
 
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=leaky
 
[upsample]
stride=2
 
[route]
layers = -1, 61
 
 
 
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=leaky
 
[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=512
activation=leaky
 
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=leaky
 
[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=512
activation=leaky
 
[convolutional]
batch_normalize=1
filters=256
size=1
stride=1
pad=1
activation=leaky
 
[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=512
activation=leaky
 
[convolutional]
size=1
stride=1
pad=1
filters=24    #filters = 3 * ( classes + 5 )   here,filters=3*(3+5)
activation=linear
 
[yolo]
mask = 3,4,5
anchors = 10,13,  16,30,  33,23,  30,61,  62,45,  59,119,  116,90,  156,198,  373,326
classes=3  #修改为自己的类别数
num=9
jitter=.3
ignore_thresh = .5
truth_thresh = 1
random=1
 
[route]
layers = -4
 
[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=leaky
 
[upsample]
stride=2
 
[route]
layers = -1, 36
 
 
 
[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=leaky
 
[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=256
activation=leaky
 
[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=leaky
 
[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=256
activation=leaky
 
[convolutional]
batch_normalize=1
filters=128
size=1
stride=1
pad=1
activation=leaky
 
[convolutional]
batch_normalize=1
size=3
stride=1
pad=1
filters=256
activation=leaky
 
[convolutional]
size=1
stride=1
pad=1
filters=24    #filters = 3 * ( classes + 5 )   here,filters=3*(3+5)
activation=linear
 
[yolo]
mask = 0,1,2
anchors = 10,13,  16,30,  33,23,  30,61,  62,45,  59,119,  116,90,  156,198,  373,326
classes=3   #修改为自己的类别数
num=9
jitter=.3
ignore_thresh = .5
truth_thresh = 1
random=1

训练模型：

./darknet detector train myData/voc.data myData/yolov3-voc.cfg darknet53.conv.74 2>&1 | tee logs/train_yolov3.log

2>&1 | tee logs/train_yolov3.log：保存训练log，如果多GPU的话，在末尾加入-gpus 0，1，2

训练完模型，测试命令

/darknet detector test myData/voc.data myData/yolov3-voc.cfg backup/yolov3-voc_20000.weights test.jpg

版权声明：本文为CSDN博主「__JDM__」的原创文章，遵循CC 4.0 BY-SA版权协议，转载请附上原文出处链接及本声明。
原文链接：https://blog.csdn.net/WSNjiang/article/details/119636685

标签： · · ·
点赞 0

__JDM__

我还没有学会写个人说明！

暂无评论

发表评论

相关推荐

【darknet】2、yolov4模型训练之模型训练
机器视觉 2024-09-23

【darknet】2、yolov4模型训练之模型训练

本文是继上一篇 数据准备之后，进行模型训练的全流程。本文直接从jupter notebook转换而来，都经过验证，但格式可能会有点问题。1、进行模型训练数据准备 使用的数据集是VOC2007_DEST

分享到：