Instructions For Use
Used to create a brand new Training Data Cleaning project, importing scattered, structured training data folders (containing images and labels) that comply with SuperMap GIS specifications, and initializing them into a manageable project file (.sda.xml).
Available starting from SuperMap iDesktopX 2026 version.
Function Entry
- Start Tab -> Browse Group -> Sample Drop-down Menu -> Training Data Cleaning -> New Project.
- After successfully creating or opening a Training Data Cleaning project -> Training Data Cleaning Tab -> Project Management -> New.
Operation Instructions
-
Click the Start Tab -> Browse Group -> Sample Drop-down Menu -> Training Data Cleaning -> New Project button. The New Project dialog box pops up.
-
Set project parameters in the dialog box:
-
Project Name: Enter the identifier name for the project.
-
Training Data Path: Click the selection button on the right to browse and select the training data folder generated by SuperMap GIS series products in the local file folder. This folder needs to contain a compliant structure (such as Images and Masks subdirectories, and *.sda files).
-
Storage Path: Specify a local directory for saving the project file (.sda.xml) and process data generated during editing.
-
- Click the OK button to automatically load the data and create the project.
After a successful new project, all imported training data is categorized under the No Category group by default in the Training Data Manager. Users need to manually classify them into corresponding groups such as All Positive Samples, All Negative Samples, Modified, etc., to complete Training Data Cleaning. For detailed operations, please refer to Samples Category And Label Editing.
Notice:The project file (.sda.xml) only saves the data path, operation records, and configuration information, and does not contain the original images and label data themselves. Do not move or delete the source data under the Training Data Path arbitrarily, as this may cause the project to fail to open or result in data loss .
Related Topics
Samples Category And Label Editing