BlendingToolKit
Tools to create blend catalogs, produce training samples and implement blending metrics
Workflow
- Step 1: import raw catalogs from CatSim or the DC2 catalogs
- Step 2: modify those catalogs to create catalogs of (multiple) blends and explore joint distributions (separation, magnitude difference)
- Step 3: generate PSF-convolved images and perform data augmentation
- Step 4: train deep-learning algorthms to do various blending-related tasks such as detection, segmentation, deblending and measurements
- Step 5: test algorithms with a set of low- and high-level performance metrics