CV领域数据集汇总

1、COCO数据集

person
bicycle
car
motorbike
aeroplane
bus
train
truck
boat
traffic light
fire hydrant
stop sign
parking meter
bench
bird
cat
dog
horse
sheep
cow
elephant
bear
zebra
giraffe
backpack
umbrella
handbag
tie
suitcase
frisbee
skis
snowboard
sports ball
kite
baseball bat
baseball glove
skateboard
surfboard
tennis racket
bottle
wine glass
cup
fork
knife
spoon
bowl
banana
apple
sandwich
orange
broccoli
carrot
hot dog
pizza
donut
cake
chair
sofa
pottedplant
bed
diningtable
toilet
tvmonitor
laptop
mouse
remote
keyboard
cell phone
microwave
oven
toaster
sink
refrigerator
book
clock
vase
scissors
teddy bear
hair drier
toothbrush

2、VOC数据集类别名

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

3、Visdrone 数据集和训练模型

3.1数据集简介

txt格式，每行从左往右分别：
,,,,,,,

YOLOv5训练visdrone数据集模型：
https://download.csdn.net/download/weixin_51154380/76852338
https://download.csdn.net/download/weixin_51154380/61665581
YOLOv3训练visdrone数据集模型：
https://download.csdn.net/download/weixin_51154380/62900703
Darknet版YOLOv3训练Visdrone数据集模型：
https://download.csdn.net/download/weixin_51154380/85356551
Darknet版YOLOv4训练Visdrone数据集模型:
https://download.csdn.net/download/weixin_51154380/85356535

3.2 检测结果可视化

3.3 类别名

pedestrian
people
bicycle
car
van
truck
tricycle
awning-tricycle
bus
motor

[ 'pedestrian', 'people', 'bicycle', 'car', 'van', 'truck', 'tricycle', 'awning-tricycle', 'bus', 'motor' ]

id2cls = {
    0: 'pedestrian',
    1: 'people',
    2: 'bicycle',
    3: 'car',
    4: 'van',
    5: 'truck',
    6: 'tricycle',
    7: 'awning-tricycle',
    8: 'bus',
    9: 'motor'
}

cls2id = {
    'pedestrian': 0,
    'people': 1,
    'bicycle': 2,
    'car': 3,
    'van': 4,
    'truck': 5,
    'tricycle': 6,
    'awning-tricycle': 7,
    'bus': 8,
    'motor': 9
}

3.4 Visdrone数据集将标签变为YOLO格式

# _*_ coding:utf-8 _*_
from tqdm import tqdm
import cv2
import time
import os
"""
Visdrone_txt -> YOLO_txt
时间： 2022.4.29
作者： 余小晖
"""

def convert_box(size, box):
    # Convert VisDrone box to YOLO xywh box
    # size= (w,h)
    dw = 1. / size[0]
    dh = 1. / size[1]
    xywh = (box[0] + box[2] / 2) * dw, (box[1] + box[3] / 2) * dh, box[2] * dw, box[3] * dh
    return xywh


file_path = './Visdrone/'
# files = os.listdir(file_path)
files = [ 'VisDrone2019-DET-train','VisDrone2019-DET-train', 'VisDrone2019-DET-val']
for file in files:
    img_path = file_path + file + '/images/'
    ann_path = file_path + file + '/annotations/'
    ann_save_path = file_path + file + '/labels-1/'
    if not os.path.exists(ann_save_path):
        os.makedirs(ann_save_path)
    imgs = os.listdir(img_path)
    for i, img in enumerate(tqdm(imgs)):
        if i == 9999:
            a = 1
        totle_name = img.split('.')[0]
        ann_name = totle_name+ '.txt'
        image = cv2.imread(img_path+ img)
        img_size = image.shape
        size = (img_size[1], img_size[0])
        with open(ann_path+ ann_name, 'r') as f:
            lines = f.readlines()
            for line in lines:
                line = line.strip().split(',')
                cls = int(line[5]) - 1  # Visdrone数据集原始标签中ID从1开始计数
                if line[4] != '0':  # line[4]=0 表示该数据有问题？
                    box = [float(x) for x in line]
                    xywh = convert_box(size, box)   # convert  VisDrone box to YOLO xywh box
                    label_value = [str(cls), str(xywh[0]), str(xywh[1]), str(xywh[2]), str(xywh[3])]
                    label_value = str.join(' ', label_value)
                    with open(ann_save_path+ ann_name, 'a')as s:
                        s.write(label_value+ '\n')
                else:
                    continue
    print(time.ctime())

4、无人机数据集和训练模型

各种尺度无人机数据集，classes: drone ; 数据量：数万张；标签格式：VOC和YOLO格式；用于无人机的视觉检测和跟踪

数据集：
https://download.csdn.net/download/weixin_51154380/32037320
https://download.csdn.net/download/weixin_51154380/33247430
https://download.csdn.net/download/weixin_51154380/33505984
https://download.csdn.net/download/weixin_51154380/40682385
https://download.csdn.net/download/weixin_51154380/41115718
https://download.csdn.net/download/weixin_51154380/70898609
https://download.csdn.net/download/weixin_51154380/85152808
https://download.csdn.net/download/weixin_51154380/85152887
https://download.csdn.net/download/weixin_51154380/85152886

Darknet版YOLOv3、v4无人机检测模型：
https://download.csdn.net/download/weixin_51154380/85351292
https://download.csdn.net/download/weixin_51154380/51112089
Pytorch版YOLOv5无人机模型：
https://download.csdn.net/download/weixin_51154380/51087322
YOLOv5-Deepsort无人机视觉检测和跟踪：
https://download.csdn.net/download/weixin_51154380/61540776

5、吸烟检测

YOLOv5吸烟检测模型下载

检测结果可视化：

6、行人检测

数据集1
数据集2
YOLOv5行人检测训练权重

检测结果可视化：

7、自行车检测数据集

1、标签格式为VOC和YOLO两种格式
2、classes: bike
2、 数量： 1800
下载链接

8、火焰和烟雾检测数据集

1、标签格式为VOC和YOLO
2、classes: fire、smoke
2、 数量： 1360+
下载链接

9、水果检测数据集

• VOC和YOLO格式标签
• 类别：香蕉、苹果、橘子、草莓
• 数量： 700+
• 下载链接

10、行人摔倒检测数据集

• 标签格式为VOC和YOLO两种格式
• classes: fall
• 数量： 8500
• 下载链接