Images
In this tutorial we will focus on working with images using Supervisely SDK.
You will learn how to:
git clone https://github.com/supervisely-ecosystem/tutorial-image.git
cd tutorial-image
./create_venv.sh
Step 3. Open repository directory in Visual Studio Code.
code -r .
Step 4. Change workspace ID in
local.env file by copying the ID from the context menu of the workspace.WORKSPACE_ID=654 # ⬅️ change value
Step 5. Start debugging
src/main.py.import os
from dotenv import load_dotenv
import supervisely as sly
First, we load environment variables with credentials and init API for communicating with Supervisely Instance.
if sly.is_development():
load_dotenv("local.env")
load_dotenv(os.path.expanduser("~/supervisely.env"))
api = sly.Api()
In this tutorial, you will need an workspace ID that you can get from environment variables. Learn more here
workspace_id = sly.env.workspace_id()
Create new project.
Source code:
project = api.project.create(workspace_id, "Fruits", change_name_if_conflict=True)
print(f"Project ID: {project.id}")
Output:
# Project ID: 15599
Create new dataset.
Source code:
dataset = api.dataset.create(project.id, "Fruits ds1")
print(f"Dataset ID: {dataset.id}")
Output:
# Dataset ID: 53465
Source code:
original_dir = "src/images/original"
path = os.path.join(original_dir, "lemons.jpg")
meta = {"my-field-1": "my-value-1", "my-field-2": "my-value-2"}
image = api.image.upload_path(
dataset.id,
name="Lemons",
path=path,
meta=meta # optional
)
print(f'Image "{image.name}" uploaded to Supervisely with ID:{image.id}')
Output:
# Image "Lemons.jpeg" uploaded to Supervisely platform with ID:17539453
✅ Supervisely API allows uploading multiple images in a single request. The code sample below sends fewer requests and it leads to a significant speed-up of our original code.
Source code:
names = [
"grapes-1.jpg",
"grapes-2.jpg",
"oranges-2.jpg",
"oranges-1.jpg",
]
paths = [os.path.join(original_dir, name) for name in names]
upload_info = api.image.upload_paths(dataset.id, names, paths)
print(f"{len(upload_info)} images successfully uploaded to Supervisely platform")
Output:
# 4 images successfully uploaded to Supervisely platform
Source code:
img_np = sly.image.read(path)
np_image_info = api.image.upload_np(dataset.id, name="Lemons-np.jpeg", img=img_np)
print(f"Image successfully uploaded as NumPy matrix to Supervisely (ID: {np_image_info.id})")
Output:
# Image successfully uploaded as NumPy matrix to Supervisely (ID: 17539458)
Source code:
names_np = [f"np-{name}" for name in names]
images_np = [sly.image.read(img_path) for img_path in paths]
np_images_info = api.image.upload_nps(dataset.id, names_np, images_np)
print(f"{len(images_np)} images successfully uploaded to platform as NumPy matrix")
Output:
# 4 images successfully uploaded to platform as NumPy matrix
Get information about image from Supervisely by id.
Source code:
image_info = api.image.get_info_by_id(image.id)
print(image_info)
Output:
# ImageInfo(
# id=17539453,
# name='Lemons.jpeg',
# link=None,
# hash='0jirgXvKGTJ8Yi0I9nCdf9MllQ9jP3Les1fD7/dt+Zk=',
# mime='image/jpeg',
# ext='jpeg',
# size=66066,
# width=640,
# height=427,
# labels_count=0,
# dataset_id=54008,
# created_at='2022-12-23T15:57:35.707Z',
# updated_at='2022-12-23T15:57:35.707Z',
# meta={},
# path_original='nAXBxaxQJRARr0Ljkj6FfREj1Fq89.jpg',
# full_storage_url='https://dev.supervise.ly/h5unublic/images/original/3/e/OK/d8Y7NnEj1Fq89.jpg',
# tags=[]
# )
You can also get information about image from Supervisely by name.
Source code:
image_name = get_file_name(image.name)
image_info_by_name = api.image.get_info_by_name(dataset.id, image_name)
print(f"image name - {image_info_by_name.name}")
Output:
# image name - Lemons.jpeg
Get information about image from Supervisely by id.
Source code:
image_info_list = api.image.get_list(dataset.id)
print(f"{len(image_info_list)} images information received.")
Output:
# 10 images information received.
Download image from Supervisely to local directory by id.
Source code:
save_path = os.path.join(result_dir, image_info.name)
api.image.download_path(image_info.id, save_path)
print(f"Image has been successfully downloaded to '{save_path}'")
Output:
# Image has been successfully downloaded to 'src/images/result/Lemons.jpeg'
Download list of images from Supervisely to local directory by ids.
Source code:
image_ids = [img.id for img in image_info_list]
image_names = [img.name for img in image_info_list]
save_paths = [os.path.join(result_dir, img_name) for img_name in image_names]
api.image.download_paths(dataset.id, image_ids, save_paths)
print(f"{len(image_info_list)} images has been successfully downloaded.")
Output:
# 10 images has been successfully downloaded.
Download image from Supervisely to local directory by id.
Source code:
image_np = api.image.download_np(image_info.id)
print(f"Image downloaded as RGB NumPy matrix. Image shape: {image_np.shape}")
Output:
# Image downloaded as RGB NumPy matrix. Image shape: (427, 640, 3)
Download list of images from Supervisely to local directory by ids.
Source code:
image_np = api.image.download_nps(dataset.id, image_ids)
print(f"{len(image_np)} images downloaded in RGB NumPy matrix.")
Output:
# 10 images downloaded in RGB NumPy matrix.
Source code:
image_info = api.image.get_info_by_id(image.id)
meta = image_info.meta
print(meta)
Output:
# {'my-field-1': 'my-value-1', 'my-field-2': 'my-value-2'}
Source code:
new_meta = {"my-field-3": "my-value-3", "my-field-4": "my-value-4"}
new_image_info = api.image.update_meta(id=image.id, meta=new_meta)
print(new_image_info["meta"])
Output:
# {'my-field-3': 'my-value-3', 'my-field-4': 'my-value-4'}
You can also get image metadata in Image labeling toolbox interface
Remove image from Supervisely by id
Source code:
api.image.remove(image.id)
print(f"Image (ID: {image.id}) successfully removed")
Output:
# Image (ID: 17539453) successfully removed
Remove list of images from Supervisely by ids.
Source code:
images_to_remove = api.image.get_list(dataset.id)
remove_ids = [img.id for img in images_to_remove]
api.image.remove_batch(remove_ids)
print(f"{len(remove_ids)} images successfully removed.")
Output:
# 9 images successfully removed.
Last modified 2mo ago