Prompt Segmentation
Feature Description
Prompt segmentation based on the SAM model utilizes vector data as prompts to segment imagery. SuperMap currently provides three segmentation capabilities: no-prompt, rectangular box prompt, and polygon prompt, enabling precise target segmentation in imagery.
Parameter Description
| Parameter Name | Default Value | Description | Type |
|---|---|---|---|
| Dataset | Select the target imagery dataset in the workspace. | DatasetImage | |
| Prompt Type | No Prompt | Supports three types: "Rectangular Box Prompt", "Polygon Prompt", and "No Prompt Data". Default is "No Prompt". | String |
| Prompt Data | Supports vector region datasets. | DatasetVector | |
| Model Magnitude | 4 | The larger the parameter quantity, the more complex the model, and generally the higher the accuracy of the segmentation result. | String |
| Sampling Method | Supports equidistant sampling and local similarity sampling. When prompt data type is "No Prompt", defaults to equidistant sampling; when prompt data is "Polygon Prompt", defaults to local similarity sampling; not selectable when prompt data type is "Rectangular Box Prompt". | Enum | |
| Tile Size | 1024 | Image dimension input to the model during inference. Default is 1024. | Integer |
| Tile Overlap (Pixels) | 0 | Default is 0, reducing insufficient prediction at tile edges. Larger overlap increases inference time. | Integer |
| Batch Size | 1 | Number of images processed per inference step. Default is 1. Within reasonable limits, larger batch sizes increase memory (VRAM) usage but reduce inference time. | Integer |
| Processor Type | Use Central Processing Unit (CPU) or Graphics Processing Unit (GPU) for computation. GPUs offer higher computational speed. | String | |
| GPU ID | 0 | Specifies GPU device ID for processing. Default is 0. Use commas to separate multiple GPUs. Becomes -1 (unselectable) when using CPU. | String |
| Result Datasource | Specifies the datasource storing the result dataset. | DataSource | |
| Result Dataset | resultDataName | Outputs vector dataset as segmentation result. | String |
Output
| Parameter Name | Description | Type |
|---|---|---|
| Prompt Segmentation Result | Output vector dataset | DatasetVector |