Binarize image object (:binaryzation)

Declaration

img = img:binaryzation({
    [white_background = white_background,]
    [csim_mode = similarity_mode,]
    [csim_algorithm = color_similarity_algorithm,]
    {color*, tolerance*},
    {color*, tolerance*},
    ...
})

img = img:binaryzation("cx*-cox*,cx*-cox*...")

Parameters

• white_background 20250302
  Boolean, optional. Whether to set the background to white. Default false (white foreground on black background)

• similarity_mode 20250302
  Boolean, optional. Whether to use similarity mode. In similarity mode, tolerance* means similarity instead of color difference. Default false

• color_similarity_algorithm 20250302
  Integer, optional. Only takes effect in similarity mode. Default 0

  Color similarity algorithms

  0 XXT default algorithm, steep gradient, similarity drops quickly with difference
  1 Manhattan distance, smooth gradient
  2 Euclidean distance, smooth gradient

• color*, tolerance*
  Integer. Whitelist of colors. color* is the color value; tolerance* is the max deviation for that color

• cx*-cox*
  String. Whitelist of colors. cx* is the hex string of color value; cox* is the hex string of max deviation for cx*

Returns

• img
  Image object. Returns the image object itself after binarization

Description

Binarize an image object
Mutates the image in-place
Performance: no data copy during operation

Examples

local pic = screen.image(462, 242, 569, 272)
pic = pic:binaryzation({
    {0x9D5D39, 0x0F1F26},
    {0xD3D3D2, 0x2C2C2D},
})

local pic = screen.image(462, 242, 569, 272)
pic = pic:binaryzation("9D5D39-0F1F26,D3D3D2-2C2C2D")

local pic = screen.image(462, 242, 569, 272)
pic = pic:binaryzation({
    csim_mode = true,        -- use similarity mode
    csim_algorithm = 2,      -- use Euclidean algorithm
    white_background = true, -- set background to white
    {0x9D5D39, 90},          -- similarity 90% for color 0x9D5D39
    {0xD3D3D2, 90},          -- similarity 90% for color 0xD3D3D2
})