text_sam module¶
The LangSAM model for segmenting objects from satellite images using text prompts. The source code is adapted from the https://github.com/luca-medeiros/lang-segment-anything repository. Credits to Luca Medeiros for the original implementation.
LangSAM
¶
A Language-based Segment-Anything Model (LangSAM) class which combines GroundingDINO and SAM.
Source code in samgeo/text_sam.py
121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 |
|
__init__(model_type='vit_h', checkpoint=None)
¶
Initialize the LangSAM instance.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
model_type
|
str
|
The model type. It can be one of the SAM 1 models () vit_h, vit_l, vit_b) or SAM 2 models (sam2-hiera-tiny, sam2-hiera-small, sam2-hiera-base-plus, sam2-hiera-large) Defaults to 'vit_h'. See https://bit.ly/3VrpxUh for more details. |
'vit_h'
|
checkpoint_url
|
str
|
The URL to the checkpoint file. Defaults to None |
required |
Source code in samgeo/text_sam.py
126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 |
|
build_groundingdino()
¶
Build the GroundingDINO model.
Source code in samgeo/text_sam.py
181 182 183 184 185 186 187 188 |
|
build_sam(model_type, checkpoint_url=None)
¶
Build the SAM model.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
model_type
|
str
|
The model type. It can be one of the SAM 1 models () vit_h, vit_l, vit_b) or SAM 2 models (sam2-hiera-tiny, sam2-hiera-small, sam2-hiera-base-plus, sam2-hiera-large) Defaults to 'vit_h'. See https://bit.ly/3VrpxUh for more details. |
required |
checkpoint_url
|
str
|
The URL to the checkpoint file. Defaults to None |
None
|
Source code in samgeo/text_sam.py
149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 |
|
predict(image, text_prompt, box_threshold, text_threshold, output=None, mask_multiplier=255, dtype=np.uint8, save_args={}, return_results=False, return_coords=False, detection_filter=None, **kwargs)
¶
Run both GroundingDINO and SAM model prediction.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
image
|
Image
|
Input PIL Image. |
required |
text_prompt
|
str
|
Text prompt for the model. |
required |
box_threshold
|
float
|
Box threshold for the prediction. |
required |
text_threshold
|
float
|
Text threshold for the prediction. |
required |
output
|
str
|
Output path for the prediction. Defaults to None. |
None
|
mask_multiplier
|
int
|
Mask multiplier for the prediction. Defaults to 255. |
255
|
dtype
|
dtype
|
Data type for the prediction. Defaults to np.uint8. |
uint8
|
save_args
|
dict
|
Save arguments for the prediction. Defaults to {}. |
{}
|
return_results
|
bool
|
Whether to return the results. Defaults to False. |
False
|
detection_filter
|
callable
|
Callable with box, mask, logit, phrase, and index args returns a boolean. If provided, the function will be called for each detected object. Defaults to None. |
None
|
Returns:
Name | Type | Description |
---|---|---|
tuple |
Tuple containing masks, boxes, phrases, and logits. |
Source code in samgeo/text_sam.py
275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 |
|
predict_batch(images, out_dir, text_prompt, box_threshold, text_threshold, mask_multiplier=255, dtype=np.uint8, save_args={}, merge=True, verbose=True, **kwargs)
¶
Run both GroundingDINO and SAM model prediction for a batch of images.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
images
|
list
|
List of input PIL Images. |
required |
out_dir
|
str
|
Output directory for the prediction. |
required |
text_prompt
|
str
|
Text prompt for the model. |
required |
box_threshold
|
float
|
Box threshold for the prediction. |
required |
text_threshold
|
float
|
Text threshold for the prediction. |
required |
mask_multiplier
|
int
|
Mask multiplier for the prediction. Defaults to 255. |
255
|
dtype
|
dtype
|
Data type for the prediction. Defaults to np.uint8. |
uint8
|
save_args
|
dict
|
Save arguments for the prediction. Defaults to {}. |
{}
|
merge
|
bool
|
Whether to merge the predictions into a single GeoTIFF file. Defaults to True. |
True
|
Source code in samgeo/text_sam.py
426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 |
|
predict_dino(image, text_prompt, box_threshold, text_threshold)
¶
Run the GroundingDINO model prediction.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
image
|
Image
|
Input PIL Image. |
required |
text_prompt
|
str
|
Text prompt for the model. |
required |
box_threshold
|
float
|
Box threshold for the prediction. |
required |
text_threshold
|
float
|
Text threshold for the prediction. |
required |
Returns:
Name | Type | Description |
---|---|---|
tuple |
Tuple containing boxes, logits, and phrases. |
Source code in samgeo/text_sam.py
190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 |
|
predict_sam(image, boxes)
¶
Run the SAM model prediction.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
image
|
Image
|
Input PIL Image. |
required |
boxes
|
Tensor
|
Tensor of bounding boxes. |
required |
Returns:
Type | Description |
---|---|
Masks tensor. |
Source code in samgeo/text_sam.py
218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 |
|
raster_to_vector(image, output, simplify_tolerance=None, **kwargs)
¶
Save the result to a vector file.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
image
|
str
|
The path to the image file. |
required |
output
|
str
|
The path to the vector file. |
required |
simplify_tolerance
|
float
|
The maximum allowed geometry displacement. The higher this value, the smaller the number of vertices in the resulting geometry. |
None
|
Source code in samgeo/text_sam.py
593 594 595 596 597 598 599 600 601 602 603 |
|
region_groups(image, connectivity=1, min_size=10, max_size=None, threshold=None, properties=None, intensity_image=None, out_csv=None, out_vector=None, out_image=None, **kwargs)
¶
Segment regions in an image and filter them based on size.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
image
|
Union[str, DataArray, ndarray]
|
Input image, can be a file path, xarray DataArray, or numpy array. |
required |
connectivity
|
int
|
Connectivity for labeling. Defaults to 1 for 4-connectivity. Use 2 for 8-connectivity. |
1
|
min_size
|
int
|
Minimum size of regions to keep. Defaults to 10. |
10
|
max_size
|
Optional[int]
|
Maximum size of regions to keep. Defaults to None. |
None
|
threshold
|
Optional[int]
|
Threshold for filling holes. Defaults to None, which is equal to min_size. |
None
|
properties
|
Optional[List[str]]
|
List of properties to measure. See https://scikit-image.org/docs/stable/api/skimage.measure.html#skimage.measure.regionprops Defaults to None. |
None
|
intensity_image
|
Optional[Union[str, DataArray, ndarray]]
|
Intensity image to use for properties. Defaults to None. |
None
|
out_csv
|
Optional[str]
|
Path to save the properties as a CSV file. Defaults to None. |
None
|
out_vector
|
Optional[str]
|
Path to save the vector file. Defaults to None. |
None
|
out_image
|
Optional[str]
|
Path to save the output image. Defaults to None. |
None
|
Returns:
Type | Description |
---|---|
Union[Tuple[ndarray, DataFrame], Tuple[DataArray, DataFrame]]
|
Union[Tuple[np.ndarray, pd.DataFrame], Tuple[xr.DataArray, pd.DataFrame]]: Labeled image and properties DataFrame. |
Source code in samgeo/text_sam.py
617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 |
|
save_boxes(output=None, dst_crs='EPSG:4326', **kwargs)
¶
Save the bounding boxes to a vector file.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
output
|
str
|
The path to the output vector file. |
None
|
dst_crs
|
str
|
The destination CRS. Defaults to "EPSG:4326". |
'EPSG:4326'
|
**kwargs
|
Additional arguments for boxes_to_vector(). |
{}
|
Source code in samgeo/text_sam.py
492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 |
|
set_image(image)
¶
Set the input image.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
image
|
str
|
The path to the image file or a HTTP URL. |
required |
Source code in samgeo/text_sam.py
257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 |
|
show_anns(figsize=(12, 10), axis='off', cmap='viridis', alpha=0.4, add_boxes=True, box_color='r', box_linewidth=1, title=None, output=None, blend=True, **kwargs)
¶
Show the annotations (objects with random color) on the input image.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
figsize
|
tuple
|
The figure size. Defaults to (12, 10). |
(12, 10)
|
axis
|
str
|
Whether to show the axis. Defaults to "off". |
'off'
|
cmap
|
str
|
The colormap for the annotations. Defaults to "viridis". |
'viridis'
|
alpha
|
float
|
The alpha value for the annotations. Defaults to 0.4. |
0.4
|
add_boxes
|
bool
|
Whether to show the bounding boxes. Defaults to True. |
True
|
box_color
|
str
|
The color for the bounding boxes. Defaults to "r". |
'r'
|
box_linewidth
|
int
|
The line width for the bounding boxes. Defaults to 1. |
1
|
title
|
str
|
The title for the image. Defaults to None. |
None
|
output
|
str
|
The path to the output image. Defaults to None. |
None
|
blend
|
bool
|
Whether to show the input image. Defaults to True. |
True
|
kwargs
|
dict
|
Additional arguments for matplotlib.pyplot.savefig(). |
{}
|
Source code in samgeo/text_sam.py
512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 |
|
show_map(basemap='SATELLITE', out_dir=None, **kwargs)
¶
Show the interactive map.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
basemap
|
str
|
The basemap. It can be one of the following: SATELLITE, ROADMAP, TERRAIN, HYBRID. |
'SATELLITE'
|
out_dir
|
str
|
The path to the output directory. Defaults to None. |
None
|
Returns:
Type | Description |
---|---|
leafmap.Map: The map object. |
Source code in samgeo/text_sam.py
605 606 607 608 609 610 611 612 613 614 615 |
|
load_model_hf(repo_id, filename, ckpt_config_filename, device='cpu')
¶
Loads a model from HuggingFace Model Hub.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
repo_id
|
str
|
Repository ID on HuggingFace Model Hub. |
required |
filename
|
str
|
Name of the model file in the repository. |
required |
ckpt_config_filename
|
str
|
Name of the config file for the model in the repository. |
required |
device
|
str
|
Device to load the model onto. Default is 'cpu'. |
'cpu'
|
Returns:
Type | Description |
---|---|
Module
|
torch.nn.Module: The loaded model. |
Source code in samgeo/text_sam.py
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 |
|
transform_image(image)
¶
Transforms an image using standard transformations for image-based models.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
image
|
Image
|
The PIL Image to be transformed. |
required |
Returns:
Type | Description |
---|---|
Tensor
|
torch.Tensor: The transformed image as a tensor. |
Source code in samgeo/text_sam.py
99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 |
|