2.0 KiB
2.0 KiB
Image Tile Merge
Node Function: The Image Tile Merge node intelligently merges multiple image tiles back into a complete image using an advanced weight mask system and cosine function gradient algorithm, ensuring perfect seamless stitching effects in merged images.
Inputs
| Parameter Name | Input Selection | Data Type | Default Value | Value Range | Description |
|---|---|---|---|---|---|
tiles |
Required | IMAGE | - | - | Image tile batch to be merged |
tile_meta |
Required | DICT | - | - | Tile metadata from ImageTileSplit node |
blend_strength |
- | FLOAT | 1.0 | 0.0-1.0 | Blend strength controlling overlap area blending degree using cosine function gradient algorithm |
Outputs
| Output Name | Data Type | Description |
|---|---|---|
merged_image |
IMAGE | Complete merged image |
Function Description
Core Technical Features
- Weight Mask System: Uses weight map accumulation for perfect blending in overlap areas
- Cosine Function Gradient: Employs mathematically smoother cosine functions for natural transition effects
- High-Precision Calculation: Uses 64-bit floating point for accumulation, avoiding precision loss and numerical errors
- Smart Boundary Handling: Automatically handles edge tile special cases ensuring complete coverage
Optimal Configuration Recommendations
- Recommended Settings:
blend_strength = 1.0(default value) - Effect: Achieves completely seamless image stitching suitable for all types of image processing tasks
Application Scenarios
- AI processing result merging: Merge AI-processed image tiles back into complete image
- Super-resolution reconstruction: Merge super-resolution processed image tiles
- Large image processing: Reconstruct large-sized images after tile processing
- Batch image stitching: High-quality image stitching and panoramic composition
- Lossless image reconstruction: Ensure processed images perfectly match original images