86 lines
2.8 KiB
Markdown
86 lines
2.8 KiB
Markdown
|
|
# Documentation
|
|
- Class name: OtsuThreshold
|
|
- Category: Bmad/CV/Thresholding
|
|
- Output node: False
|
|
|
|
OtsuThreshold节点使用Otsu阈值法对图像进行前景和背景的分割。它可以选择性地使用高斯模糊对图像进行预处理，以减少噪声并改善阈值分割的结果。
|
|
|
|
# Input types
|
|
## Required
|
|
- image
|
|
- 需要进行阈值分割的输入图像。这是Otsu方法将要应用的主要数据。
|
|
- Comfy dtype: IMAGE
|
|
- Python dtype: torch.Tensor
|
|
- threshold_type
|
|
- 指定要应用的阈值分割类型，这会影响前景和背景的区分方式。
|
|
- Comfy dtype: COMBO[STRING]
|
|
- Python dtype: str
|
|
- gaussian_blur_x
|
|
- 高斯模糊的核宽度。较高的值意味着更强的模糊效果，这有助于在阈值分割前减少噪声。
|
|
- Comfy dtype: INT
|
|
- Python dtype: int
|
|
- gaussian_blur_y
|
|
- 高斯模糊的核高度。与gaussian_blur_x一起定义模糊的程度。
|
|
- Comfy dtype: INT
|
|
- Python dtype: int
|
|
- gaussian_border_type
|
|
- 决定在高斯模糊过程中如何处理图像的边界。
|
|
- Comfy dtype: COMBO[STRING]
|
|
- Python dtype: str
|
|
|
|
# Output types
|
|
- image
|
|
- 应用Otsu阈值分割后的输出图像，其中前景和背景已被分离。
|
|
- Comfy dtype: IMAGE
|
|
- Python dtype: torch.Tensor
|
|
|
|
|
|
## Usage tips
|
|
- Infra type: `GPU`
|
|
- Common nodes: unknown
|
|
|
|
|
|
## Source code
|
|
```python
|
|
class OtsuThreshold:
|
|
|
|
@classmethod
|
|
def INPUT_TYPES(s):
|
|
return {
|
|
"required": {
|
|
"image": ("IMAGE",),
|
|
# "channel": (s.channels, {"default": "greyscale"}),
|
|
"threshold_type": (thresh_types, {"default": thresh_types[0]}),
|
|
"gaussian_blur_x": ("INT", {
|
|
"default": 4,
|
|
"min": 0,
|
|
"max": 200,
|
|
"step": 2
|
|
}),
|
|
"gaussian_blur_y": ("INT", {
|
|
"default": 4,
|
|
"min": 0,
|
|
"max": 200,
|
|
"step": 2
|
|
}),
|
|
"gaussian_border_type": (border_types, {"default": border_types[0]}),
|
|
},
|
|
}
|
|
|
|
RETURN_TYPES = ("IMAGE",)
|
|
FUNCTION = "otsu_thresthold"
|
|
CATEGORY = "Bmad/CV/Thresholding"
|
|
|
|
def otsu_thresthold(self, image, threshold_type, gaussian_blur_x, gaussian_blur_y, gaussian_border_type):
|
|
image = tensor2opencv(image, 1)
|
|
if gaussian_blur_x > 0 and gaussian_blur_y > 0:
|
|
image = cv.GaussianBlur(image, (gaussian_blur_x + 1, gaussian_blur_y + 1),
|
|
border_types_map[gaussian_border_type])
|
|
_, image = cv.threshold(image, 0, 255, thresh_types_map[threshold_type] + cv.THRESH_OTSU)
|
|
image = cv.cvtColor(image, cv.COLOR_GRAY2RGB)
|
|
image = opencv2tensor(image)
|
|
return (image,)
|
|
|
|
```
|