Feature Description
Change detection compares multi-temporal images of the same area (typically collected over different periods) to determine the type, magnitude, and location of changes. It holds significant and extensive application value in natural resource domains such as land use, urban expansion, geological hazard monitoring, ecological environment protection, and forest conservation.
SuperMap iDesktopX provides general change detection functionality to serve various image change detection scenarios, identifying changed land parcels and generating vector results of changed pixels in target categories. This feature enables fine-grained image interpretation, typically used to detect prominent feature changes like buildings, roads, and water bodies.
Parameter Description
| Parameter Name | Default Value | Description | Type |
|---|---|---|---|
| File Type | Dataset | Specify file type. Select between dataset or image list | String |
| Source Data - Datasource | Specify input data source. Supports file database | String | |
| Source Data - Dataset | Specify input image dataset | String | |
| Source Data - List Path | Specify input image list path. Supports txt and csv formats. Each line in the list corresponds to an image file path, matching line-by-line with compare data list | String | |
| Compare Data - Datasource | Specify comparison image data source. Supports file database | String | |
| Compare Data - Dataset | Specify comparison image dataset | String | |
| Compare Data - List Path | Specify comparison image list path. Supports txt and csv formats. Each line corresponds to a comparison image file path, matching line-by-line with source data list | String | |
| Model File | Specify model file for change detection | String | |
| Tile Overlap (Pixels) | 0 | Specify tile overlap value (pixels) to reduce insufficient prediction at tile edges. Larger overlap increases inference time. Recommended not to exceed 1/4 of model training data size (check tileSize field in sdm model) | Integer |
| Batch Size | 1 | Specify number of images processed per inference step | Integer |
| Processor Type | GPU | Specify processor type | String |
| GPU ID | 0 | Input GPU ID. For multi-GPU inference, input multiple IDs separated by commas (e.g., "0,1,2") | String |
| Other Parameter Settings (Optional) |
false | Check to enable bounds dataset/datasource parameter configuration | Boolean |
| Result Type | Vector Dataset | Select to output result as vector dataset or raster file | String |
| Result Datasource | Specify datasource to save vector dataset | String | |
| Result Dataset | CD_result | Specify vector dataset name. For batch processing, original image names will be appended with "_" to distinguish results | String |
| Result Save Path | Specify folder path to save raster files | String | |
| File Name | CD_result | Specify raster file name. For batch processing, original image names will be appended with "_" to distinguish results | String |
Output
| Parameter Name | Description | Type |
|---|---|---|
| Detected Result | Vector dataset or raster file obtained through general change detection | String |