Darknet实现YoloV3（1）

环境：环境 ubantu16.04+cudnn7.0+cuda_9.0.176

1、安装darknet

 $ git clone https://github.com/pjreddie/darknet 
 $ cd darknet

2、修改Makefile

GPU=1 #0或1 使用GPU为1，不使用为0。
CUDNN=1 #0或1
OPENCV=1 #0或1
OPENMP=0
DEBUG=0

3、编译

$ make

4、下载预训练模型

$ wget https://pjreddie.com/media/files/yolov3.weights 

5、用预训练模型进行简单的测试

1. 一张图片测试：

$ ./darknet detect cfg/yolov3.cfg yolov3.weights data/dog.jpg 

2)视频检测：

$ ./darknet detector demo cfg/coco.data cfg/yolov3.cfg yolov3.weights person.mp4

6、训练VOC风格的数据

所用数据集为COCO中抽取的person类转化为VOC风格。
darknet不需要xml文件，需要.txt文件，用voc_label.py生成
修改voc_label.py文件
修改sets为训练样本集的名称
sets=[(‘2007’, ‘train’),(‘2007’,‘val’),(‘2007’,‘test’)]
修改classes为训练样本集的类标签
classes=[“person”]
在voc_label.py文件中加上os.system(“cat 2007_train.txt 2007_val.txt > train.txt”)
或直接在voc_label.py所在文件夹中打开终端执行cat 2007_train.txt 2007_val.txt > train.txt命令。

7、下载Imagenet上预先训练的权重

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

8、修改cfg/voc.data

classes= 1 #classes为训练样本集的类别总数
train = /home/user/darknet/train.txt #train的路径为训练样本集所在的路径
valid = /home/user/darknet/2007_test.txt #valid的路径为验证样本集所在的路径
names = data/voc.names #names的路径为data/voc.names文件所在的路径
backup = backup

9、修改data/voc.name为样本集的标签名

在文件中写入person

10、修改cfg/yolov3-voc.cfg

[net]
# Testing            ### 测试模式                                          
# batch=1
# subdivisions=1
# Training           ### 训练模式，每次前向的图片数目 = batch/subdivisions
batch=64
subdivisions=16
width=416            ### 网络的输入宽、高、通道数
height=416
channels=3
momentum=0.9         ### 动量
decay=0.0005         ### 权重衰减
angle=0
saturation = 1.5     ### 饱和度
exposure = 1.5       ### 曝光度
hue=.1               ### 色调
learning_rate=0.001  ### 学习率
burn_in=1000         ### 学习率控制的参数
max_batches = 50200  ### 迭代次数       我设置的是10000                                   
policy=steps         ### 学习率策略
steps=40000,45000    ### 学习率变动步长   我设置的是8500,9500
scales=.1,.1         ### 学习率变动因子  

[convolutional]
batch_normalize=1    ### BN
filters=32           ### 卷积核数目
size=3               ### 卷积核尺寸
stride=1             ### 卷积核步长
pad=1                ### pad
activation=leaky     ### 激活函数
......
[convolutional]
size=1
stride=1
pad=1
filters=18  #3*(1+4+1)
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=1  #类别
num=9
jitter=.3
ignore_thresh = .5
truth_thresh = 1
random=1  #1，如果显存很小，将random设置为0，关闭多尺度训练；
......
[convolutional]
size=1
stride=1
pad=1
filters=18  #3*(1+4+1)
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=1  #类别
num=9
jitter=.3
ignore_thresh = .5
truth_thresh = 1
random=1  #1，如果显存很小，将random设置为0，关闭多尺度训练；
......
[convolutional]
size=1
stride=1
pad=1
filters=18  #3*(1+4+1)
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=1  #类别
num=9
jitter=.3  # 数据扩充的抖动操作
ignore_thresh = .5  #文章中的阈值1
truth_thresh = 1  #文章中的阈值2
random=1 #1，如果显存很小，将random设置为0，关闭多尺度训练；

11、开始训练

1）运行命令：

$ ./darknet detector train cfg/voc.data cfg/yolov3-voc.cfg darknet53.conv.74

训练过程产生的文件保存在backup文件夹中。
yolov3-voc.backup可以继续训练

 $ ./darknet detector train cfg/voc.data cfg/yolov3-voc.cfg backup/yolov3-voc.backup  

2）训练日志可视化
前提是训练过程中保存了训练日志xxx.log.
①运行测试时的命令改为：

$ ./darknet detector train cfg/voc.data cfg/yolov3-voc.cfg weights/darknet53.conv.74 -gpu 0 | tee train_yolov3.log

训练结束后会生成train_yolov3.log脚本

② 在使用脚本绘制变化曲线之前，需要先使用extract_log.py脚本，格式化log,用生成的新的log文件供可视化工具绘图，格式化log的extract_log.py脚本如下（和生成的log文件同一目录）参考：（https://blog.csdn.net/qq_34806812/article/details/81459982）
extract_log.py：

# coding=utf-8
# 该文件用来提取训练log，去除不可解析的log后使log文件格式化，生成新的log文件供可视化工具绘图
import inspect
import os
import random
import sys
def extract_log(log_file,new_log_file,key_word):
    with open(log_file, 'r') as f:
      with open(new_log_file, 'w') as train_log:
  #f = open(log_file)
    #train_log = open(new_log_file, 'w')
        for line in f:
    # 去除多gpu的同步log
          if 'Syncing' in line:
            continue
    # 去除除零错误的log
          if 'nan' in line:
            continue
          if key_word in line:
            train_log.write(line)
    f.close()
    train_log.close()
extract_log('train_yolov3.log','train_log_loss.txt','images')
extract_log('train_yolov3.log','train_log_iou.txt','IOU')

train_log_loss.txt内容：

train_log_iou.txt内容：

③在运行下面脚本：参考（https://blog.csdn.net/csdn_zhishui/article/details/85397380）

# -*- coding: utf-8 -*-
# @Time    : 2018/12/30 16:26
# @Author  : lazerliu
# @File    : vis_yolov3_log.py
# @Func    :yolov3 训练日志可视化，把该脚本和日志文件放在同一目录下运行。

import pandas as pd
import matplotlib.pyplot as plt
import os

# ==================可能需要修改的地方=====================================#
g_log_path = "train_yolov3.log"  # 此处修改为你的训练日志文件名
# ==========================================================================#

def extract_log(log_file, new_log_file, key_word):
    '''
    :param log_file:日志文件
    :param new_log_file:挑选出可用信息的日志文件
    :param key_word:根据关键词提取日志信息
    :return:
    '''
    with open(log_file, "r") as f:
        with open(new_log_file, "w") as train_log:
            for line in f:
                # 去除多gpu的同步log
                if "Syncing" in line:
                    continue
                # 去除nan log
                if "nan" in line:
                    continue
                if key_word in line:
                    train_log.write(line)
    f.close()
    train_log.close()


def drawAvgLoss(loss_log_path):
    '''
    :param loss_log_path: 提取到的loss日志信息文件
    :return: 画loss曲线图
    '''
    line_cnt = 0
    for count, line in enumerate(open(loss_log_path, "rU")):
        line_cnt += 1
    result = pd.read_csv(loss_log_path, skiprows=[iter_num for iter_num in range(line_cnt) if ((iter_num < 500))],
                         error_bad_lines=False,
                         names=["loss", "avg", "rate", "seconds", "images"])
    result["avg"] = result["avg"].str.split(" ").str.get(1)
    result["avg"] = pd.to_numeric(result["avg"])

    fig = plt.figure(1, figsize=(6, 4))
    ax = fig.add_subplot(1, 1, 1)
    ax.plot(result["avg"].values, label="Avg Loss", color="#ff7043")
    ax.legend(loc="best")
    ax.set_title("Avg Loss Curve")
    ax.set_xlabel("Batches")
    ax.set_ylabel("Avg Loss")


def drawIOU(iou_log_path):
    '''
    :param iou_log_path: 提取到的iou日志信息文件
    :return: 画iou曲线图
    '''
    line_cnt = 0
    for count, line in enumerate(open(iou_log_path, "rU")):
        line_cnt += 1
    result = pd.read_csv(iou_log_path, skiprows=[x for x in range(line_cnt) if (x % 39 != 0 | (x < 5000))],
                         error_bad_lines=False,
                         names=["Region Avg IOU", "Class", "Obj", "No Obj", "Avg Recall", "count"])
    result["Region Avg IOU"] = result["Region Avg IOU"].str.split(": ").str.get(1)

    result["Region Avg IOU"] = pd.to_numeric(result["Region Avg IOU"])

    result_iou = result["Region Avg IOU"].values
    # 平滑iou曲线
    for i in range(len(result_iou) - 1):
        iou = result_iou[i]
        iou_next = result_iou[i + 1]
        if abs(iou - iou_next) > 0.2:
            result_iou[i] = (iou + iou_next) / 2

    fig = plt.figure(2, figsize=(6, 4))
    ax = fig.add_subplot(1, 1, 1)
    ax.plot(result_iou, label="Region Avg IOU", color="#ff7043")
    ax.legend(loc="best")
    ax.set_title("Avg IOU Curve")
    ax.set_xlabel("Batches")
    ax.set_ylabel("Avg IOU")

if __name__ == "__main__":
    loss_log_path = "train_log_loss.txt"
    iou_log_path = "train_log_iou.txt"
    if os.path.exists(g_log_path) is False:
        exit(-1)
    if os.path.exists(loss_log_path) is False:
        extract_log(g_log_path, loss_log_path, "images")
    if os.path.exists(iou_log_path) is False:
        extract_log(g_log_path, iou_log_path, "IOU")
    drawAvgLoss(loss_log_path)
    drawIOU(iou_log_path)
    plt.show()

12、测试单张图片：

测试时的cfg文件中的batch和subdivisions必须为1。
运行命令：

$ ./darknet detector test cfg/voc.data cfg/yolov3-voc.cfg backup/yolov3-voc_final.weights timg.jpeg

13、批量测试：

参考：https://blog.csdn.net/wangzy9766/article/details/88749696
测试时：Makefile中的opencv调为0。
1）用下面代码替换detector.c文件（example文件夹下）的void test_detector函数（注意有3处要改成自己的路径）
全部复制并代替，三处修改路径写对
此段代码来自https://blog.csdn.net/mieleizhi0522/article/details/79989754

void test_detector(char *datacfg, char *cfgfile, char *weightfile, char *filename, float thresh, float hier_thresh, char *outfile, int fullscreen)
{
    list *options = read_data_cfg(datacfg);
    char *name_list = option_find_str(options, "names", "data/names.list");
    char **names = get_labels(name_list);

    image **alphabet = load_alphabet();
    network *net = load_network(cfgfile, weightfile, 0);
    set_batch_network(net, 1);
    srand(2222222);
    double time;
    char buff[256];
    char *input = buff;
    float nms=.45;
    int i=0;
    while(1){
        if(filename){
            strncpy(input, filename, 256);
            image im = load_image_color(input,0,0);
            image sized = letterbox_image(im, net->w, net->h);
        //image sized = resize_image(im, net->w, net->h);
        //image sized2 = resize_max(im, net->w);
        //image sized = crop_image(sized2, -((net->w - sized2.w)/2), -((net->h - sized2.h)/2), net->w, net->h);
        //resize_network(net, sized.w, sized.h);
            layer l = net->layers[net->n-1];


            float *X = sized.data;
            time=what_time_is_it_now();
            network_predict(net, X);
            printf("%s: Predicted in %f seconds.\n", input, what_time_is_it_now()-time);
            int nboxes = 0;
            detection *dets = get_network_boxes(net, im.w, im.h, thresh, hier_thresh, 0, 1, &nboxes);
            //printf("%d\n", nboxes);
            //if (nms) do_nms_obj(boxes, probs, l.w*l.h*l.n, l.classes, nms);
            if (nms) do_nms_sort(dets, nboxes, l.classes, nms);
                draw_detections(im, dets, nboxes, thresh, names, alphabet, l.classes);
                free_detections(dets, nboxes);
            if(outfile)
             {
                save_image(im, outfile);
             }
            else{
                save_image(im, "predictions");
#ifdef OPENCV
                cvNamedWindow("predictions", CV_WINDOW_NORMAL);
                if(fullscreen){
                cvSetWindowProperty("predictions", CV_WND_PROP_FULLSCREEN, CV_WINDOW_FULLSCREEN);
                }
                show_image(im, "predictions");
                cvWaitKey(0);
                cvDestroyAllWindows();
#endif
            }
            free_image(im);
            free_image(sized);
            if (filename) break;
         }
        else {
            printf("Enter Image Path: ");
            fflush(stdout);
            input = fgets(input, 256, stdin);
            if(!input) return;
            strtok(input, "\n");
   
            list *plist = get_paths(input);
            char **paths = (char **)list_to_array(plist);
             printf("Start Testing!\n");
            int m = plist->size;
            if(access("/home/FENGsl/darknet/data/out",0)==-1)//"/home/FENGsl/darknet/data"修改成自己的路径
            {
              if (mkdir("/home/FENGsl/darknet/data/out",0777))//"/home/FENGsl/darknet/data"修改成自己的路径
               {
                 printf("creat file bag failed!!!");
               }
            }
            for(i = 0; i < m; ++i){
             char *path = paths[i];
             image im = load_image_color(path,0,0);
             image sized = letterbox_image(im, net->w, net->h);
        //image sized = resize_image(im, net->w, net->h);
        //image sized2 = resize_max(im, net->w);
        //image sized = crop_image(sized2, -((net->w - sized2.w)/2), -((net->h - sized2.h)/2), net->w, net->h);
        //resize_network(net, sized.w, sized.h);
        layer l = net->layers[net->n-1];


        float *X = sized.data;
        time=what_time_is_it_now();
        network_predict(net, X);
        printf("Try Very Hard:");
        printf("%s: Predicted in %f seconds.\n", path, what_time_is_it_now()-time);
        int nboxes = 0;
        detection *dets = get_network_boxes(net, im.w, im.h, thresh, hier_thresh, 0, 1, &nboxes);
        //printf("%d\n", nboxes);
        //if (nms) do_nms_obj(boxes, probs, l.w*l.h*l.n, l.classes, nms);
        if (nms) do_nms_sort(dets, nboxes, l.classes, nms);
        draw_detections(im, dets, nboxes, thresh, names, alphabet, l.classes);
        free_detections(dets, nboxes);
        if(outfile){
            save_image(im, outfile);
        }
        else{
             
             char b[2048];
            sprintf(b,"/home/FENGsl/darknet/data/out/%s",GetFilename(path));//"/home/FENGsl/darknet/data"修改成自己的路径
            
            save_image(im, b);
            printf("save %s successfully!\n",GetFilename(path));
#ifdef OPENCV
            cvNamedWindow("predictions", CV_WINDOW_NORMAL);
            if(fullscreen){
                cvSetWindowProperty("predictions", CV_WND_PROP_FULLSCREEN, CV_WINDOW_FULLSCREEN);
            }
            show_image(im, "predictions");
            cvWaitKey(0);
            cvDestroyAllWindows();
#endif
        }

        free_image(im);
        free_image(sized);
        if (filename) break;
        }
      }
    }
}

2）.在前面添加GetFilename(char p)函数（注意后面的注释）
全部复制（包括头文件）
此段代码来自https://blog.csdn.net/mieleizhi0522/article/details/79989754

#include "darknet.h"
#include <sys/stat.h>
#include<stdio.h>
#include<time.h>
#include<sys/types.h>
static int coco_ids[] = {1,2,3,4,5,6,7,8,9,10,11,13,14,15,16,17,18,19,20,21,22,23,24,25,27,28,31,32,33,34,35,36,37,38,39,40,41,42,43,44,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,67,70,72,73,74,75,76,77,78,79,80,81,82,84,85,86,87,88,89,90};

char *GetFilename(char *p)
{
    static char name[20]={""};
    char *q = strrchr(p,'/') + 1;
    strncpy(name,q,6);//注意后面的6，如果你的测试集的图片的名字字符（不包括后缀）是其他长度，请改为你需要的长度（官方的默认的长度是6）
    return name;
}

3）重新进行编译

$ make clean
$ make

4）读取测试图片的准备
如果要另找图片进行测试，不用2007_test.txt进行测试需要:
①新建要测试图像的文件夹
②制作自己的txt：
dataset.py程序：

import os
path = ['/home/yuxin/darknet']
f = open('input_data.txt','w')
for path in path:
   p = os.path.abspath(path) + '/pictures_person'
   filenames = os.listdir(p)
   for filename in filenames:
      im_path = p + '/' + filename
      print(im_path)
      f.write(im_path + '\n')
f.close()

input_data.txt中的内容如下:

5）执行批量测试命令
./darknet detector test cfg/voc.data cfg/yolov3-voc.cfg backup/yolov3-voc_final.weights

6)输入Image Path:

14、生成预测结果：

执行命令：

$ ./darknet detector valid cfg/voc.data cfg/yolov3-voc.cfg backup/yolov3-voc_final.weights -out“”

结果生成在<data_cfg>的指定的目录下以<out_file>开头的若干文件中，若<data_cfg>没有指定results，那么默认为<darknet_root>/results

默认情况下，结果会得到在results下的各类别的txt检测结果文件
每行代表一个预测框，分别为：不带路径与后缀的图片名，置信度分数，四个绝对坐标值
此处的-out后面直接" "即可，因为不论你写什么，他都会根据voc_names中配置的类名来生成对应txt，有几个类就生成几个txt，并且会将类名自动写入txt文件名中。
person.txt内容：

15、计算mAP

采用第三方工具计算mAP
下载第三方库：

$ git clone https://github.com/LianjiLi/yolo-compute-map.git

voc_eval.py程序如下（使用faster RCNN的voc_eval进行计算，只是把最后的返回值改成return ap）：

voc_eval.py程序如下（使用faster RCNN的voc_eval进行计算，只是把最后的返回值改成return ap）：

compute_mAP.py程序如下：
计算单类时：

from voc_eval import voc_eval
print voc_eval('/home/yuxin/darknet/results/{}.txt', '/home/yuxin/darknet/VOCdevkit/VOC2007/Annotations/{}.xml', '/home/yuxin/darknet/VOCdevkit/VOC2007/ImageSets/Main/test.txt', 'person', '.')

计算多类时：

from voc_eval import voc_eval

map_ = 0
classnames = ['person']  #填写自己的类别
for classname in classnames:
    ap = voc_eval('/home/yuxin/darknet/results/{}.txt', '/home/yuxin/darknet/VOCdevkit/VOC2007/Annotations/{}.xml', '/home/yuxin/darknet/VOCdevkit/VOC2007/ImageSets/Main/test.txt', classname, '.')
    map_ += ap
    print ('%-20s' % (classname + '_ap:')+'%s' % ap)

map = map_/len(classnames)
print ('%-20s' % 'map:' + '%s' % map)

或者：（ https://blog.csdn.net/m0_37857151/article/details/86605087）

from voc_eval import voc_eval

import os
current_path = os.getcwd()
results_path = current_path+"/results"
sub_files = os.listdir(results_path)
mAP = []
for i in range(len(sub_files)):
    class_name = sub_files[i].split(".txt")[0]
    rec, prec, ap = voc_eval('/home/yuxin/darknet/results/{}.txt', '/home/yuxin/darknet/VOCdevkit/VOC2007/Annotations/{}.xml', '/home/yuxin/darknet/VOCdevkit/VOC2007/ImageSets/Main/test.txt', class_name, '.')
    print("{} :\t {} ".format(class_name, ap))
    mAP.append(ap)
mAP = tuple(mAP)
print("***************************")
print("mAP :\t {}".format( float( sum(mAP)/len(mAP)) ))

上述程序在python2下运行：

16、训练之后计算召回率：

1） 修改detector.c文件中validate_detector_recall函数
list *plist = get_paths(“data/coco_val_5k.list”);
换成：
list *plist=get_paths("/home/…/darknet/train.txt");

2）for(k = 0; k < l.wl.hl.n; ++k){

换成：
for(k = 0; k < nboxes; ++k){
重新 make
3）执行命令：

$ ./darknet detector recall cfg/voc.data cfg/yolov3-voc.cfg backup/yolov3-voc_final.weights

参考博客：
https://www.cnblogs.com/pprp/p/9525508.html
https://blog.csdn.net/wangzy9766/article/details/88749696
https://www.cnblogs.com/xieqi/p/9818056.html
https://blog.csdn.net/m0_37857151/article/details/86605087
https://blog.csdn.net/csdn_zhishui/article/details/85397380