使用PIL从EXIF数据中获取照片拍摄的日期和时间

Question

我可以通过使用PIL库来获取图片的EXIF数据，但是我该怎么获取照片拍摄的日期和时间呢？

Answer 1

我创建了一个函数（get_exif），可以更方便地获取一些属性。

使用方法

your_date_time: str = get_exif(Path('test.jpg'), 'DateTimeOriginal')
# your_date_time = get_exif(Path('test.jpg'), 0x9003)  # same as above

get_exif

from pathlib import Path
from typing import Union, List, Tuple
import PIL.Image
import PIL.ExifTags


def get_exif(file_path: Path,
             search_list: Union[int, str, List, Tuple] = None,
             ignore_error=True
             ) -> Union[int, PIL.Image.Exif, List]:
    """
    :param file_path: image path
    :param search_list: if you want to get some property, then you can pass the id or name, it will return by order.
    :param ignore_error:
    :return:
        int: -1 FileNotFoundError, or exif is None
        PIL.Image.Exif: when the `search_list` is None, return the whole Exif
    """
    tag_by_id: dict = PIL.ExifTags.TAGS
    try:
        im: PIL.Image.Image = PIL.Image.open(str(file_path))
    except FileNotFoundError:
        if ignore_error:
            return -1
        else:
            raise FileNotFoundError(file_path)
    exif: PIL.Image.Exif = im.getexif()
    if not exif:
        if ignore_error:
            return -1
        else:
            raise ValueError("exif is None")
    if search_list is None:
        return exif
    tag_by_name = {tag_by_id[dec_value]: exif[dec_value] for dec_value in exif if dec_value in tag_by_id}
    result_list = []
    if not isinstance(search_list, (list, tuple)):
        search_list = [search_list]
    for key in search_list:
        if isinstance(key, int):
            result_list.append(exif.get(key, None))
            continue
        try:
            dec_value = int(key, 16)
            result_list.append(exif.get(dec_value, None))
            continue
        except ValueError:
            ...
        result_list.append(tag_by_name.get(key, None))
    return result_list if len(result_list) > 1 else result_list[0]

更多测试

import unittest

class ExifTests(unittest.TestCase):
    def test(self):
        exif: PIL.Image.Exif = get_exif(Path('test.jpg'))  # same as the ``PIL.Image.open().getexif()``

        # get specify attribute only
        date_t_ori1 = get_exif(Path('test.jpg'), 0x9003)
        date_t_ori2 = get_exif(Path('test.jpg'), '0x9003', )  # hex string is ok too.
        date_t_ori3 = get_exif(Path('test.jpg'), 'DateTimeOriginal')  # Give name is also working.
        self.assertTrue(date_t_ori1 == date_t_ori2 == date_t_ori3)

        # You can get multiple values at once. If the key does not exist, it returns None.
        date_t_ori4, img_length, _, __ = get_exif(Path('test.jpg'),
                                                  (36867, 'ImageLength', 'NoteExitName', -12345))

        # Occurring error return -1 (by default ignore_error=True)
        self.assertEqual(-1, get_exif(Path('not exist.jpg')))  # FileNotFoundError
        self.assertEqual(-1, get_exif(Path('no_exif_tag.jpg')))  # ValueError

        self.assertRaises(FileNotFoundError, get_exif, Path('not exist.jpg'), ignore_error=False)
        self.assertRaises(ValueError, get_exif, Path('no_exif_tag.jpg'), ignore_error=False)

注意

这个 网站 很不错，但如果你想要获取完整的列表，建议查看 PIL.ExifTags.py，请看下面的内容。

# ExifTags.py

# Maps EXIF tags to tag names.

TAGS = {
    # possibly incomplete
    0x000B: "ProcessingSoftware",
    0x00FE: "NewSubfileType",
    0x00FF: "SubfileType",
    0x0100: "ImageWidth",
    0x0101: "ImageLength",
    0x0102: "BitsPerSample",
    ...
    0xA430: "CameraOwnerName",  # <-- The website does not record it. (The website record last tag is A420.)
    0xA431: "BodySerialNumber",
    0xA432: "LensSpecification",
    0xA433: "LensMake",
    0xA434: "LensModel",
}


# Maps EXIF GPS tags to tag names.

GPSTAGS = {
    0: "GPSVersionID",
    1: "GPSLatitudeRef",
    2: "GPSLatitude",
    3: "GPSLongitudeRef",
    4: "GPSLongitude",
    5: "GPSAltitudeRef",
    6: "GPSAltitude",
    7: "GPSTimeStamp",
    ...
}

Answer 2

ExifTags.TAGS 是一个标签和标签名称之间的对应关系。你可以用它来创建一个标签名称和对应值的映射。

在这张特定的图片上，有几个不同的“日期”属性（DateTime、DateTimeOriginal、DateTimeDigitized）可以使用。

import json
from PIL import Image, ExifTags
from datetime import datetime

def main(filename):
    image_exif = Image.open(filename)._getexif()
    if image_exif:
        # Make a map with tag names
        exif = { ExifTags.TAGS[k]: v for k, v in image_exif.items() if k in ExifTags.TAGS and type(v) is not bytes }
        print(json.dumps(exif, indent=4))
        # Grab the date
        date_obj = datetime.strptime(exif['DateTimeOriginal'], '%Y:%m:%d %H:%M:%S')
        print(date_obj)
    else:
        print('Unable to get date from exif for %s' % filename)

输出：

{
    "DateTimeOriginal": "2008:11:15 19:36:24",
    "DateTimeDigitized": "2008:11:15 19:36:24",
    "ColorSpace": 1,
    "ExifImageWidth": 3088,
    "SceneCaptureType": 0,
    "ExifImageHeight": 2320,
    "SubjectDistanceRange": 2,
    "ImageDescription": "               ",
    "Make": "Hewlett-Packard                ",
    "Model": "HP Photosmart R740             ",
    "Orientation": 1,
    "DateTime": "2008:11:15 19:36:24",
    ...
}
2008-11-15 19:36:24

Answer 3

在最近的Pillow版本中，这个情况有了一些变化（我记得是6.0以上的版本）。

他们新增了一个公开的方法getexif()，你应该使用这个方法。之前的版本是一个私有的、实验性的功能（_getexif()）。

from PIL import Image

im = Image.open('path/to/image.jpg')
exif = im.getexif()
creation_time = exif.get(36867)

Answer 4

我喜欢使用 exif-py，因为它是纯Python写的，不需要编译或安装，而且可以在Python 2.x和3.x上运行，非常适合和小型便携的Python应用程序一起打包。

链接： https://github.com/ianare/exif-py

下面是一个获取照片拍摄日期和时间的例子：

import exifread
with open('image.jpg', 'rb') as fh:
    tags = exifread.process_file(fh, stop_tag="EXIF DateTimeOriginal")
    dateTaken = tags["EXIF DateTimeOriginal"]
    return dateTaken

Answer 5

最后我找到了答案，我需要的标签是 36867：

from PIL import Image
def get_date_taken(path):
    exif = Image.open(path)._getexif()
    if not exif:
        raise Exception('Image {0} does not have EXIF data.'.format(path))
    return exif[36867]

这个在我使用的旧版“PIL”和新版“Pillow”上都能正常工作。