Instructions for use
Binary classification uses spectral information and geospatial data of various ground objects in remote sensing images, and employs deep learning algorithms to classify a specific type of ground object element in remote sensing images. It determines whether the pixels in the image are of the interest category and generates binary raster data with values of 0 and 1, where 0 represents non-interest categories and 1 represents the interest category.
The binary classification function interprets images at the pixel level with fine granularity. The raster results are highly flexible and can be optimized through a series of post-processing steps. The resulting raster data can be converted into vector data for further analysis.
Binary classification is typically used to obtain single land cover types with distinct features, such as roads, rivers, and buildings. Information such as the location, boundaries, and area of the interest category can be calculated through pixels.
Parameter description
-
Source Dataset：Select the image that requires binary classification. It supports two types: datasets and folders.
-
-
If the file type is set to dataset, it supports input file-based image data (such as *.tif, *.img, etc.), image datasets, and mosaic datasets.
-
If the file type is set to folder, you need to input the image folder path.
-
Notes:- Only supports 8-bit unsigned format images.
- Supports three-band and multi-band images (images with 4 or more bands), and the band order of the image must be consistent with the sample image used during model training.
-
Model File：Select Model File (*.sdm).
-
Tile Overlap (Pixels)：During binary classification, due to the structural characteristics of convolutional neural networks, data at the edges of tiles may not be fully predicted. To improve prediction quality, tile overlap is required. This parameter is in pixels and can be set to one-eighth to one-quarter of the tile size. The larger the tile overlap in pixels, the longer it takes to infer the entire image. The tile size corresponding to the model is recorded in the tile_size parameter of the model file (*.sdm). For example, if the tile size is 1024, the tile overlap in pixels can be set to 128-256.
-
Batch Size：Due to the large volume of remote sensing image data and limitations in computer performance, the model cannot be loaded all at once. Therefore, during inference, a block-by-block reading and processing method is used. The batch size refers to the number of slices inferred simultaneously. Appropriately increasing the batch size can improve interpretation efficiency, but it is limited by the video memory (for GPU inference) or memory (for CPU inference) size of the inference device.
-
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 the processor type is set to CPU, this value is fixed at -1. If using GPU, you need to query the GPU ID via 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 identification numbers for data processing. To use multiple GPUs, separate them with English commas ",", such as "0,2,3", indicating the use of GPUs with IDs 0, 2, and 3.
-
Other Parameters：Check other parameters to input a region dataset with the same coordinate system as the source data, used to limit the inference scope.
-
Result Data：The result of binary classification can be output as a vector dataset or raster file.
-
If selecting vector dataset, you need to specify the data source for saving the vector dataset and the vector dataset name. For batch inference, the original image name will be appended with an underscore "_" to the vector dataset name to distinguish binary classification results from multiple images.
-
If selecting raster file, you need to specify the folder path for saving the raster file and the raster file name. For batch inference, the original image name will be appended with an underscore "_" to the raster file name to distinguish binary classification results from multiple images.
-
Related topics