darknet-yolov3训练自己的数据

比眉伴天荒 2023-02-23 05:09 26阅读 0赞

注意：本篇博客直接使用VOC2007数据集

1.数据集

Labelimg软件构建数据集，Labelimg项目地址[：https://github.com/tzutalin/labelImg，][https_github.com_tzutalin_labelImg]Labelimg快捷键：

<table> 
 <tbody> 
  <tr> 
   <td>Ctrl + u</td> 
   <td>Load all of the images from a directory</td> 
  </tr> 
  <tr> 
   <td>Ctrl + r</td> 
   <td>Change the default annotation target dir</td> 
  </tr> 
  <tr> 
   <td>Ctrl + s</td> 
   <td>Save</td> 
  </tr> 
  <tr> 
   <td>Ctrl + d</td> 
   <td>Copy the current label and rect box</td> 
  </tr> 
  <tr> 
   <td>Space</td> 
   <td>Flag the current image as verified</td> 
  </tr> 
  <tr> 
   <td>w</td> 
   <td>Create a rect box</td> 
  </tr> 
  <tr> 
   <td>d</td> 
   <td>Next image</td> 
  </tr> 
  <tr> 
   <td>a</td> 
   <td>Previous image</td> 
  </tr> 
  <tr> 
   <td>del</td> 
   <td>Delete the selected rect box</td> 
  </tr> 
  <tr> 
   <td>Ctrl++</td> 
   <td>Zoom in</td> 
  </tr> 
  <tr> 
   <td>Ctrl--</td> 
   <td>Zoom out</td> 
  </tr> 
  <tr> 
   <td>↑→↓←</td> 
   <td>Keyboard arrows to move selected rect box</td> 
  </tr> 
 </tbody> 
</table>

voc2007数据集目录结构 ：

----voc2007

----Annotations

----ImageSets

----Main

----JPEGImages

在voc2007同目录下新建makeTXT.py，将数据集划分，并且在Main文件夹下构建4个TXT：train.txt，test.txt，trainval.txt，val.txt。代码如下：

import os
    import random
    
    trainval_percent = 0.8
    train_percent = 0.8
    xmlfilepath = 'VOC2007\Annotations'
    txtsavepath = 'VOC2007\ImageSets\Main'
    total_xml = os.listdir(xmlfilepath)
    
    num = len(total_xml)
    list = range(num)
    tv = int(num * trainval_percent)
    tr = int(tv * train_percent)
    trainval = random.sample(list, tv)
    train = random.sample(trainval, tr)
    
    ftrainval = open('VOC2007/ImageSets/Main/trainval.txt', 'w')
    ftest = open('VOC2007/ImageSets/Main/test.txt', 'w')
    ftrain = open('VOC2007/ImageSets/Main/train.txt', 'w')
    fval = open('VOC2007/ImageSets/Main/val.txt', 'w')
    
    for i in list:
        name = total_xml[i][:-4] + '\n'
        if i in trainval:
            ftrainval.write(name)
            if i in train:
                ftrain.write(name)
            else:
                fval.write(name)
        else:
            ftest.write(name)
    
    ftrainval.close()
    ftrain.close()
    fval.close()

在 voc2007同目录下新建voc\_labels.py，生成labels。代码如下：

import xml.etree.ElementTree as ET
    import pickle
    import os
    from os import listdir, getcwd
    from os.path import join
    
    #sets=[('2012', 'train'), ('2012', 'val'), ('2007', 'train'), ('2007', 'val'), ('2007', 'test')]
    
    sets=[('2007', 'train'), ('2007', 'val'), ('2007', 'test')]
    
    classes = ["aeroplane", "bicycle", "bird", "boat", "bottle", "bus", "car", "cat", "chair", "cow", "diningtable", "dog", "horse", "motorbike", "person", "pottedplant", "sheep", "sofa", "train", "tvmonitor"]
    
    
    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('VOCdevkit/VOC%s/Annotations/%s.xml'%(year, image_id))
        #out_file = open('VOCdevkit/VOC%s/labels/%s.txt'%(year, image_id), 'w')
        in_file = open('VOC%s/Annotations/%s.xml' % (year, image_id))
        out_file = open('VOC%s/labels/%s.txt'%(year, 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('VOCdevkit/VOC%s/labels/'%(year)):
            os.makedirs('VOCdevkit/VOC%s/labels/'%(year))
        image_ids = open('VOCdevkit/VOC%s/ImageSets/Main/%s.txt'%(year, image_set)).read().strip().split()
        list_file = open('%s_%s.txt'%(year, image_set), 'w')
        for image_id in image_ids:
            list_file.write('%s/VOCdevkit/VOC%s/JPEGImages/%s.jpg\n'%(wd, year, image_id))
            convert_annotation(year, image_id)
        list_file.close()
    os.system("cat 2007_train.txt 2007_val.txt 2012_train.txt 2012_val.txt > train.txt")
    os.system("cat 2007_train.txt 2007_val.txt 2007_test.txt 2012_train.txt 2012_val.txt > train.all.txt")
    '''
    for year, image_set in sets:
        if not os.path.exists('VOC%s/labels/'%(year)):
            os.makedirs('VOC%s/labels/'%(year))
        image_ids = open('VOC%s/ImageSets/Main/%s.txt'%(year, image_set)).read().strip().split()
        list_file = open('%s_%s.txt'%(year, image_set), 'w')
        for image_id in image_ids:
            list_file.write('%s/VOC%s/JPEGImages/%s.jpg\n'%(wd, year, image_id))
            convert_annotation(year, image_id)
        list_file.close()
    
    os.system("cat 2007_train.txt 2007_val.txt  > train.txt")
    os.system("cat 2007_train.txt 2007_val.txt 2007_test.txt  > train.all.txt")

2.环境

（1）git clone [https://github.com/AlexeyAB/darknet][https_github.com_AlexeyAB_darknet]

（2）cd darknet

（3）pip install -r requirements.txt

（4）make

（5）在项目根目录下新建weights文件夹，下载权重文件，将其放入weights文件夹中。

（6）测试：./darknet detect cfg/yolov3.cfg weights/yolov3.weights data/dog.jpg    或     ./darknet detector test cfg/coco.data cfg/yolov3.cfg yolov3.weights data/dog.jpg

3.训练模型

（1）下载darknet53.conv.74，将darknet53.conv.74放入其中。

（2）在data目录下新建\*\*.name文件，存放你的数据集类别名称。本文用coco.names:

aeroplane
    bicycle
    bird
    boat
    bottle
    bus
    car
    cat
    chair
    cow
    diningtable
    dog
    horse
    motorbike
    person
    pottedplant
    sheep
    sofa
    train
    tvmonitor

（3）在data目录下新建\*\*.data文件，本文用coco.data：

classes = 20#类别数
    train = data\2007_train.txt#voc_labels.py生成的训练集的位置
    valid = data\2007_test.txt
    names = data\coco.names
    backup = backup\

（4） 更新cfg文件的classes，本文使用的classes=20。yolo上一卷积层的filters=3\*（classes+5），其中5代表的是4个坐标+1个置信度。

（5）开始训练：python ./darknet detector train cfg/voc.data cfg/yolov3-voc.cfg weights/darknet53.conv.74

注意：max\_batches = 50200 \#\#\# 迭代次数

[https_github.com_tzutalin_labelImg]: https://github.com/tzutalin/labelImg
[https_github.com_AlexeyAB_darknet]: https://github.com/AlexeyAB/darknet