Generate Image Chips¶

Install Package¶

To use the geoai-py package, ensure it is installed in your environment. Uncomment the command below if needed.

In [1]:

Copied!

# %pip install geoai-py
# %pip install geoai-py

Import Packages¶

In [2]:

Copied!

import geoai
import geoai

Download sample data¶

In [3]:

Copied!





raster_url = (
    "https://huggingface.co/datasets/giswqs/geospatial/resolve/main/naip_train.tif"
)
vector_url = "https://huggingface.co/datasets/giswqs/geospatial/resolve/main/naip_train_buildings.geojson"
raster_url = (
    "https://huggingface.co/datasets/giswqs/geospatial/resolve/main/naip_train.tif"
)
vector_url = "https://huggingface.co/datasets/giswqs/geospatial/resolve/main/naip_train_buildings.geojson"

In [4]:

Copied!

raster_path = geoai.download_file(raster_url)
raster_path = geoai.download_file(raster_url)

File already exists: naip_train.tif

In [5]:

Copied!

vector_path = geoai.download_file(vector_url)
vector_path = geoai.download_file(vector_url)

naip_train_buildings.geojson:   0%|          | 0.00/456k [00:00<?, ?B/s]

naip_train_buildings.geojson: 100%|██████████| 456k/456k [00:00<00:00, 6.53MB/s]

Preview data¶

In [6]:

Copied!

geoai.view_image(raster_path, figsize=(18, 10))
geoai.view_image(raster_path, figsize=(18, 10))

No description has been provided for this image

In [7]:

Copied!

geoai.view_vector(vector_path, basemap=True, alpha=0.5, figsize=(18, 10))
geoai.view_vector(vector_path, basemap=True, alpha=0.5, figsize=(18, 10))

Out[7]:

<Axes: >

In [8]:

Copied!

geoai.view_vector_interactive(vector_path)
geoai.view_vector_interactive(vector_path)

Out[8]:

Make this Notebook Trusted to load map: File -> Trust Notebook

Convert vector to raster¶

In [9]:

Copied!

output_path = vector_path.replace(".geojson", ".tif")
geoai.vector_to_raster(vector_path, output_path, reference_raster=raster_path)
output_path = vector_path.replace(".geojson", ".tif")
geoai.vector_to_raster(vector_path, output_path, reference_raster=raster_path)

Reprojecting vector data from EPSG:4326 to EPSG:26911
Rasterized data saved to naip_train_buildings.tif

Out[9]:

array([[0, 0, 0, ..., 0, 0, 0],
       [0, 0, 0, ..., 0, 0, 0],
       [0, 0, 0, ..., 0, 0, 0],
       ...,
       [0, 0, 0, ..., 0, 0, 0],
       [0, 0, 0, ..., 0, 0, 0],
       [0, 0, 0, ..., 0, 0, 0]], shape=(1126, 2503), dtype=uint8)

In [10]:

Copied!

geoai.view_image(output_path, figsize=(18, 10))
geoai.view_image(output_path, figsize=(18, 10))

Generate image chips¶

In [11]:

Copied!





tiles = geoai.export_geotiff_tiles(
    in_raster=raster_path,
    out_folder="output",
    in_class_data=vector_path,
    tile_size=512,
    stride=256,
    buffer_radius=0,
    create_overview=True,
)
tiles = geoai.export_geotiff_tiles(
    in_raster=raster_path,
    out_folder="output",
    in_class_data=vector_path,
    tile_size=512,
    stride=256,
    buffer_radius=0,
    create_overview=True,
)

Raster info for naip_train.tif:
  CRS: EPSG:26911
  Dimensions: 2503 x 1126
  Resolution: (0.5999999999999953, 0.5999999999996691)
  Bands: 4
  Bounds: BoundingBox(left=454780.8, bottom=5277567.0, right=456282.6, top=5278242.6)
Loaded 722 features from naip_train_buildings.geojson
Vector CRS: EPSG:4326
Reprojecting features from EPSG:4326 to EPSG:26911
Found 6 unique classes: ['apartments' None 'terrace' 'detached' 'house' 'shed']

Generating tiles:   0%|          | 0/36 [00:00<?, ?it/s]

Generated: 1, With features: 1:   3%|▎         | 1/36 [00:00<00:03, 11.44it/s]

Generated: 1, With features: 1:   6%|▌         | 2/36 [00:00<00:02, 12.82it/s]

Generated: 2, With features: 2:   6%|▌         | 2/36 [00:00<00:02, 12.82it/s]

Generated: 3, With features: 3:   8%|▊         | 3/36 [00:00<00:02, 12.82it/s]

Generated: 3, With features: 3:  11%|█         | 4/36 [00:00<00:02, 10.69it/s]

Generated: 4, With features: 4:  11%|█         | 4/36 [00:00<00:02, 10.69it/s]

Generated: 5, With features: 5:  14%|█▍        | 5/36 [00:00<00:02, 10.69it/s]

Generated: 5, With features: 5:  17%|█▋        | 6/36 [00:00<00:03,  8.28it/s]

Generated: 6, With features: 6:  17%|█▋        | 6/36 [00:00<00:03,  8.28it/s]

Generated: 6, With features: 6:  19%|█▉        | 7/36 [00:00<00:03,  7.60it/s]

Generated: 7, With features: 7:  19%|█▉        | 7/36 [00:00<00:03,  7.60it/s]

Generated: 7, With features: 7:  22%|██▏       | 8/36 [00:01<00:04,  6.84it/s]

Generated: 8, With features: 8:  22%|██▏       | 8/36 [00:01<00:04,  6.84it/s]

Generated: 8, With features: 8:  25%|██▌       | 9/36 [00:01<00:03,  6.81it/s]

Generated: 9, With features: 9:  25%|██▌       | 9/36 [00:01<00:03,  6.81it/s]

Generated: 9, With features: 9:  28%|██▊       | 10/36 [00:01<00:03,  7.39it/s]

Generated: 10, With features: 10:  28%|██▊       | 10/36 [00:01<00:03,  7.39it/s]

Generated: 10, With features: 10:  31%|███       | 11/36 [00:01<00:03,  7.88it/s]

Generated: 11, With features: 11:  31%|███       | 11/36 [00:01<00:03,  7.88it/s]

Generated: 11, With features: 11:  33%|███▎      | 12/36 [00:01<00:02,  8.31it/s]

Generated: 12, With features: 12:  33%|███▎      | 12/36 [00:01<00:02,  8.31it/s]

Generated: 12, With features: 12:  36%|███▌      | 13/36 [00:01<00:02,  8.27it/s]

Generated: 13, With features: 13:  36%|███▌      | 13/36 [00:01<00:02,  8.27it/s]

Generated: 13, With features: 13:  39%|███▉      | 14/36 [00:01<00:02,  7.83it/s]

Generated: 14, With features: 14:  39%|███▉      | 14/36 [00:01<00:02,  7.83it/s]

Generated: 14, With features: 14:  42%|████▏     | 15/36 [00:01<00:03,  6.99it/s]

Generated: 15, With features: 15:  42%|████▏     | 15/36 [00:01<00:03,  6.99it/s]

Generated: 15, With features: 15:  44%|████▍     | 16/36 [00:02<00:03,  6.12it/s]

Generated: 16, With features: 16:  44%|████▍     | 16/36 [00:02<00:03,  6.12it/s]

Generated: 16, With features: 16:  47%|████▋     | 17/36 [00:02<00:03,  5.66it/s]

Generated: 17, With features: 17:  47%|████▋     | 17/36 [00:02<00:03,  5.66it/s]

Generated: 17, With features: 17:  50%|█████     | 18/36 [00:02<00:03,  5.45it/s]

Generated: 18, With features: 18:  50%|█████     | 18/36 [00:02<00:03,  5.45it/s]

Generated: 18, With features: 18:  53%|█████▎    | 19/36 [00:02<00:02,  5.95it/s]

Generated: 19, With features: 19:  53%|█████▎    | 19/36 [00:02<00:02,  5.95it/s]

Generated: 19, With features: 19:  56%|█████▌    | 20/36 [00:02<00:02,  6.23it/s]

Generated: 20, With features: 20:  56%|█████▌    | 20/36 [00:02<00:02,  6.23it/s]

Generated: 20, With features: 20:  58%|█████▊    | 21/36 [00:02<00:02,  6.40it/s]

Generated: 21, With features: 21:  58%|█████▊    | 21/36 [00:02<00:02,  6.40it/s]

Generated: 21, With features: 21:  61%|██████    | 22/36 [00:03<00:02,  6.88it/s]

Generated: 22, With features: 22:  61%|██████    | 22/36 [00:03<00:02,  6.88it/s]

Generated: 22, With features: 22:  64%|██████▍   | 23/36 [00:03<00:01,  7.25it/s]

Generated: 23, With features: 23:  64%|██████▍   | 23/36 [00:03<00:01,  7.25it/s]

Generated: 23, With features: 23:  67%|██████▋   | 24/36 [00:03<00:01,  6.43it/s]

Generated: 24, With features: 24:  67%|██████▋   | 24/36 [00:03<00:01,  6.43it/s]

Generated: 24, With features: 24:  69%|██████▉   | 25/36 [00:03<00:01,  6.41it/s]

Generated: 25, With features: 25:  69%|██████▉   | 25/36 [00:03<00:01,  6.41it/s]

Generated: 25, With features: 25:  72%|███████▏  | 26/36 [00:03<00:01,  6.04it/s]

Generated: 26, With features: 26:  72%|███████▏  | 26/36 [00:03<00:01,  6.04it/s]

Generated: 26, With features: 26:  75%|███████▌  | 27/36 [00:03<00:01,  6.01it/s]

Generated: 27, With features: 27:  75%|███████▌  | 27/36 [00:03<00:01,  6.01it/s]

Generated: 27, With features: 27:  78%|███████▊  | 28/36 [00:04<00:01,  5.96it/s]

Generated: 28, With features: 28:  78%|███████▊  | 28/36 [00:04<00:01,  5.96it/s]

Generated: 28, With features: 28:  81%|████████  | 29/36 [00:04<00:01,  5.95it/s]

Generated: 29, With features: 29:  81%|████████  | 29/36 [00:04<00:01,  5.95it/s]

Generated: 29, With features: 29:  83%|████████▎ | 30/36 [00:04<00:00,  6.50it/s]

Generated: 30, With features: 30:  83%|████████▎ | 30/36 [00:04<00:00,  6.50it/s]

Generated: 30, With features: 30:  86%|████████▌ | 31/36 [00:04<00:00,  7.03it/s]

Generated: 31, With features: 31:  86%|████████▌ | 31/36 [00:04<00:00,  7.03it/s]

Generated: 31, With features: 31:  89%|████████▉ | 32/36 [00:04<00:00,  7.06it/s]

Generated: 32, With features: 32:  89%|████████▉ | 32/36 [00:04<00:00,  7.06it/s]

Generated: 32, With features: 32:  92%|█████████▏| 33/36 [00:04<00:00,  6.69it/s]

Generated: 33, With features: 33:  92%|█████████▏| 33/36 [00:04<00:00,  6.69it/s]

Generated: 33, With features: 33:  94%|█████████▍| 34/36 [00:04<00:00,  6.30it/s]

Generated: 34, With features: 34:  94%|█████████▍| 34/36 [00:04<00:00,  6.30it/s]

Generated: 34, With features: 34:  97%|█████████▋| 35/36 [00:05<00:00,  6.26it/s]

Generated: 35, With features: 35:  97%|█████████▋| 35/36 [00:05<00:00,  6.26it/s]

Generated: 35, With features: 35: 100%|██████████| 36/36 [00:05<00:00,  5.96it/s]

Generated: 36, With features: 36: 100%|██████████| 36/36 [00:05<00:00,  5.96it/s]

Generated: 36, With features: 36: 100%|██████████| 36/36 [00:05<00:00,  6.75it/s]

Overview image saved to output/overview.png

------- Export Summary -------
Total tiles exported: 36
Tiles with features: 36 (100.0%)
Average feature pixels per tile: 45966.3
Output saved to: output

------- Georeference Verification -------

Preview image chips¶

In [12]:

Copied!

geoai.view_image("output/overview.png", figsize=(18, 10))
geoai.view_image("output/overview.png", figsize=(18, 10))

In [13]:

Copied!

geoai.view_image("output/images/tile_000000.tif")
geoai.view_image("output/images/tile_000000.tif")

In [14]:

Copied!

geoai.view_image("output/labels/tile_000000.tif")
geoai.view_image("output/labels/tile_000000.tif")

In [15]:

Copied!

geoai.view_image("output/images/tile_000001.tif")
geoai.view_image("output/images/tile_000001.tif")

In [16]:

Copied!

geoai.view_image("output/labels/tile_000001.tif")
geoai.view_image("output/labels/tile_000001.tif")