yolov4-tiny训练数据集用于gazebo平台的无人机识别（2）

参考文档：(tiny) YOLOv4 详细训练指南(附下载链接)_yolov4_tiny 预训练模型下载-CSDN博客

                   目标检测与跟踪 (yuque.com)

由于需要在gazebo上识别仿真平台提供的iris四旋翼无人机，需要对其建立数据集并训练自己的.weights文件

1.下载编译darknet并准备训练环境

2.标注数据集并调整格式

3.修改配置并开始训练

开始标注

打开labelme,选择Opendir为~/darknet/train_data/VOCdevkit/VOC2007/JPEGImages

在上方菜单选择File/Change Output Dir为~/darknet/train_data/VOCdevkit/VOC2007/Json，

开启Save Automatically

选择Edit/Create Rectangle标注框为矩形

这里我标了四种类型，包括drone、car、human、street sign，依次将所有图片完成标注

使用下方.py脚本进行转换

from __future__ import print_function

import argparse
import glob
import os
import os.path as osp
import sys

import imgviz
import labelme

try:
    import lxml.builder
    import lxml.etree
except ImportError:
    print("Please install lxml:\n\n    pip install lxml\n")
    sys.exit(1)


def main():
    parser = argparse.ArgumentParser(
        formatter_class=argparse.ArgumentDefaultsHelpFormatter
    )
    parser.add_argument("input_dir", help="input annotated directory")
    parser.add_argument("output_dir", help="output dataset directory")
    parser.add_argument("--labels", help="labels file", required=True)
    parser.add_argument(
        "--noviz", help="no visualization", action="store_true"
    )
    args = parser.parse_args()

    if osp.exists(args.output_dir):
        print("Output directory already exists:", args.output_dir)
        sys.exit(1)
    os.makedirs(args.output_dir)
    os.makedirs(osp.join(args.output_dir, "JPEGImages"))
    os.makedirs(osp.join(args.output_dir, "Annotations"))
    if not args.noviz:
        os.makedirs(osp.join(args.output_dir, "AnnotationsVisualization"))
    print("Creating dataset:", args.output_dir)

    class_names = []
    class_name_to_id = {}
    for i, line in enumerate(open(args.labels).readlines()):
        class_id = i - 1  # starts with -1
        class_name = line.strip()
        class_name_to_id[class_name] = class_id
        if class_id == -1:
            assert class_name == "__ignore__"
            continue
        elif class_id == 0:
            assert class_name == "_background_"
        class_names.append(class_name)
    class_names = tuple(class_names)
    print("class_names:", class_names)
    out_class_names_file = osp.join(args.output_dir, "class_names.txt")
    with open(out_class_names_file, "w") as f:
        f.writelines("\n".join(class_names))
    print("Saved class_names:", out_class_names_file)

    for filename in glob.glob(osp.join(args.input_dir, "*.json")):
        print("Generating dataset from:", filename)

        label_file = labelme.LabelFile(filename=filename)

        base = osp.splitext(osp.basename(filename))[0]
        out_img_file = osp.join(args.output_dir, "JPEGImages", base + ".jpg")
        out_xml_file = osp.join(args.output_dir, "Annotations", base + ".xml")
        if not args.noviz:
            out_viz_file = osp.join(
                args.output_dir, "AnnotationsVisualization", base + ".jpg"
            )

        img = labelme.utils.img_data_to_arr(label_file.imageData)
        imgviz.io.imsave(out_img_file, img)

        maker = lxml.builder.ElementMaker()
        xml = maker.annotation(
            maker.folder(),
            maker.filename(base + ".jpg"),
            maker.database(),  # e.g., The VOC2007 Database
            maker.annotation(),  # e.g., Pascal VOC2007
            maker.image(),  # e.g., flickr
            maker.size(
                maker.height(str(img.shape[0])),
                maker.width(str(img.shape[1])),
                maker.depth(str(img.shape[2])),
            ),
            maker.segmented(),
        )

        bboxes = []
        labels = []
        for shape in label_file.shapes:
            if shape["shape_type"] != "rectangle":
                print(
                    "Skipping shape: label={label}, "
                    "shape_type={shape_type}".format(**shape)
                )
                continue

            class_name = shape["label"]
            class_id = class_names.index(class_name)

            (xmin, ymin), (xmax, ymax) = shape["points"]
            # swap if min is larger than max.
            xmin, xmax = sorted([xmin, xmax])
            ymin, ymax = sorted([ymin, ymax])

            bboxes.append([ymin, xmin, ymax, xmax])
            labels.append(class_id)

            xml.append(
                maker.object(
                    maker.name(shape["label"]),
                    maker.pose(),
                    maker.truncated(),
                    maker.difficult(),
                    maker.bndbox(
                        maker.xmin(str(xmin)),
                        maker.ymin(str(ymin)),
                        maker.xmax(str(xmax)),
                        maker.ymax(str(ymax)),
                    ),
                )
            )

        if not args.noviz:
            captions = [class_names[label] for label in labels]
            viz = imgviz.instances2rgb(
                image=img,
                labels=labels,
                bboxes=bboxes,
                captions=captions,
                font_size=15,
            )
            imgviz.io.imsave(out_viz_file, viz)

        with open(out_xml_file, "wb") as f:
            f.write(lxml.etree.tostring(xml, pretty_print=True))


if __name__ == "__main__":
    main()

运行脚本

#在jsontoxml.py同级文件夹下新建labels.txt文件 
#内容为

__ignore__
_background_
drone
car
human
street sign
#根据实际标注数据类型调整

python3 jsontoxml.py --labels=labels.txt ~/darknet/train_data/VOCdevkit/VOC2007/Json ~/xmloutput
#~/darknet/train_data/VOCdevkit/VOC2007/Json 为输入文件夹 ~/xmloutput为输出文件夹

mv ~/xmloutput/Annotations/* ~/darknet/train_data/VOCdevkit/VOC2007/Annotations/
rm -r ~/xmloutput

修改~/darknet/scripts/xml2voc.py

import os
import random

trainval_percent = 1  # trainval数据集占所有数据的比例
train_percent = 0.7  # train数据集占trainval数据的比例
xmlfilepath = '../train_data/VOCdevkit/VOC2007/Annotations/'
txtsavepath = '../train_data/VOCdevkit/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(txtsavepath + 'trainval.txt', 'w')
ftest = open(txtsavepath + 'test.txt', 'w')
ftrain = open(txtsavepath + 'train.txt', 'w')
fval = open(txtsavepath + '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()
ftest .close()

python3 ~/darknet/scripts/xml2voc.py#运行生成test.txt train.txt val.txt

生成voc的label文件

将~/darknet/scripts/voc_label.py修改为以下：

import xml.etree.ElementTree as ET
import pickle
import os
from os import listdir, getcwd
from os.path import join

sets=[('2007', 'train'), ('2007', 'val'), ('2007', 'test')]

classes = ["drone", "car", "human", "street sign"]


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')
    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 (difficult is not None and 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 > train.txt")

并将其复制到train_data目录下，并运行

cp ~/darknet/scripts/voc_label.py ~/darknet/train_data/
python3 ~/darknet/train_data/voc_label.py

 最后格式如下

----train_data\
    |----2007_test.txt
    |----2007_train.txt
    |----2007_val.txt
    |----train.txt
    |----VOCdevkit\
    |    |----VOC2007\
    |    |    |----Annotations\
    |    |    |    |----00000001.xml
    |    |    |----ImageSets\
    |    |    |    |----Main\
    |    |    |    |    |----test.txt
    |    |    |    |    |----train.txt
    |    |    |    |    |----val.txt
    |    |    |----JPEGImages\
    |    |    |    |----00000001.jpg
    |    |    |----labels\
    |    |    |    |----00000001.txt
    |----voc_label.py

3.修改配置并开始训练