ControlNet - Canny Edge
Learn to use Canny edge ControlNet.
What is Canny ControlNet
Canny edge detection extracts edges from input image, providing structure guidance for generation.
Installation
Download ControlNet Model
ComfyUI/models/controlnet/
control_v11p_sd15_canny.pthDownload Preprocessor
ComfyUI usually includes preprocessors in:
ComfyUI/custom_nodes/comfyui_controlnet_aux/Workflow
[LoadImage] → [CannyEdgePreprocessor] → [ControlNet] → [Apply ControlNet]
↑
[CLIPTextEncode(positive)] → [KSampler] ← [CLIPTextEncode(negative)]Parameters
Canny Preprocessor
| Parameter | Description | Recommended |
|---|---|---|
| Low Threshold | Lower edge detection | 100-150 |
| High Threshold | Upper edge detection | 200-250 |
Lower values detect more edges (potential noise). Higher values detect only strong edges.
ControlNet Settings
| Parameter | Description | Recommended |
|---|---|---|
| Strength | Control intensity | 0.5-0.8 |
| Start | Begin control at step | 0.0 |
| End | End control at step | 1.0 |
Use Cases
Architecture
bash
# Prompt
modern building exterior, glass facade, clean design, professional photography
# Input
Floor plan or line drawingProduct Design
bash
# Prompt
sleek smartphone design, white, minimal, product photography
# Input
Product sketch or outlineCharacter Line Art
bash
# Prompt
beautiful woman, detailed face, portrait
# Input
Line art drawingTips
- Threshold adjustment: Start with 100/200, adjust based on results
- Edge quality: Use clean, clear input images
- Strength balance: Too high can flatten details
Common Issues
| Issue | Solution |
|---|---|
| Too many edges | Increase threshold |
| Too few edges | Decrease threshold |
| Style too rigid | Lower strength to 0.3-0.5 |
Next Steps
- ControlNet Depth - Depth control
- ControlNet OpenPose - Pose control
- Multi-ControlNet - Combine multiple