Instructions for use
Object extraction pre-selects regions of interest categories in images based on neural network models, then classifies pixels within the regions to obtain object boundary information, thereby supporting object-level spatial analysis.
Parameter description
-
Source Dataset: Select the image that needs object extraction, supporting both dataset and folder types.
-
-
If the file type selects dataset, it supports inputting file-based image data (such as *.tif, *.img, etc.), image dataset, and mosaic dataset.
-
If the file type selects folder, you need to input the image folder path.
-
Note:- Only supports 8-bit unsigned format images.
- Supports three-band and multi-band images (images with four or more bands), and the image band order must be consistent with the sample image used during model training.
-
Model File: Select Model File (*.sdm).
-
Probability Threshold: For each extracted object, the system calculates the probability that it matches the object features. The extraction result only retains objects with predicted probabilities higher than this value, with a default value of 0.5.
-
Deduplicate Threshold: The system generates multiple candidate boxes for extracted objects in an image and assigns probability values to each. After applying the NMS algorithm, the optimal box is obtained. Candidate boxes with an overlap greater than the deduplicate threshold with the optimal box are removed based on the Intersection over Union (IoU). The deduplicate threshold is generally 0.3-0.7, with a default value of 0.3.
-
Single-Step Computation Amount: Due to the large volume of remote sensing image data, under computer performance limitations, the model cannot be loaded all at once. Therefore, during inference, a block-by-block reading and processing method is adopted. The single-step computation amount refers to the number of slices processed simultaneously during inference. Appropriately increasing the single-step computation amount can improve interpretation efficiency, but it is limited by the inference device's VRAM (for GPU inference) or memory (for CPU inference) size.
-
Processor Type: You can use the computer's central processing unit (CPU) or graphics processing unit (GPU) to process data.
-
GPU ID: Specify the GPU ID for data processing. Default is 0. If 'Processor Type' selects CPU, this value is fixed at -1. If using GPU, you need to query the GPU ID through the 'nvidia-smi' command in the system command line. The GPU ID is based on the query result. If there are multiple GPUs, you can specify the GPU identifier for data processing. If you need to use multiple GPUs, please separate them with English commas ',', such as '0,2,3', indicating the use of GPUs with IDs 0, 2, and 3.
-
Other Parameters
- Range Data Source/Dataset: You can input a region dataset with the same coordinate system as the source data to limit the inference range.
- Return: If checked, an additional dataset of the minimum bounds of the result objects will be generated, with the dataset name ending in 'bbox'.
-
Result Data: This is the result vector dataset of object extraction. Set the data source and name for saving the result dataset. If it's batch inference, the original image name will be appended with '_' to the vector dataset name to distinguish object extraction results of multiple images.
Related Topics