如何将灰度素描图像转为彩色素描？基于OpenCV实现黑白素描配色替换的可行性及方法问询

Hey there! I've got you covered on both of your sketch color questions using OpenCV—let's break this down clearly.

问题2：将黑白线条素描改为红白/绿白等非黑配色 #

Absolutely, this is totally doable with OpenCV, and it's simpler than you might think. The core idea is leveraging the grayscale values of your sketch to separate line areas from the background, then assigning your target colors to each region.

Step-by-Step Implementation #

Let's assume your sketch is a single-channel grayscale image where lines are dark (low pixel values, ~0 for black) and the background is light (high values, ~255 for white):

import cv2
import numpy as np

# Load your grayscale sketch (use flag 0 to read as single-channel)
sketch_gray = cv2.imread("your_sketch.png", 0)
height, width = sketch_gray.shape

# ------------------- Option 1: Red lines, White background -------------------
# Create a white 3-channel base image
color_sketch = np.full((height, width, 3), 255, dtype=np.uint8)
# Define the line region (adjust threshold if your sketch isn't pure black/white)
line_mask = sketch_gray < 127
# Assign red color to lines (OpenCV uses BGR format, so red = (0, 0, 255))
color_sketch[line_mask] = (0, 0, 255)

# ------------------- Option 2: Green lines, White background -------------------
# color_sketch = np.full((height, width, 3), 255, dtype=np.uint8)
# color_sketch[line_mask] = (0, 255, 0)  # Green in BGR

# Save or preview the result
cv2.imwrite("red_white_sketch.png", color_sketch)
cv2.imshow("Custom Color Sketch", color_sketch)
cv2.waitKey(0)
cv2.destroyAllWindows()

Key Notes #

• Adjust the threshold (127 in the code) if your sketch has gray transitions instead of pure black/white—test values like 100 or 150 to get clean lines.
• For reverse color schemes (e.g., white lines on red background), just swap the base color and line color: create a red base image, then set white color to the non-line regions (sketch_gray >= 127).

问题1：将灰度素描转换为彩色素描 #

There are two common approaches here, depending on the type of color effect you want:

Approach 1: Pseudocolor Gradient Sketches #

If you want automatic, gradient-based color (like thermal map effects), use OpenCV's built-in color mapping function:

import cv2

sketch_gray = cv2.imread("gray_sketch.png", 0)
# Apply a pre-defined color map (try COLORMAP_VIRIDIS, COLORMAP_MAGMA, etc.)
color_sketch = cv2.applyColorMap(sketch_gray, cv2.COLORMAP_JET)
cv2.imwrite("pseudocolor_sketch.png", color_sketch)

Approach 2: Natural Color Sketches (Using Original Image) #

If you want lines to match the colors of the original photo (like a colored pencil sketch), combine your sketch with the original image:

import cv2
import numpy as np

original_img = cv2.imread("original_photo.jpg")
sketch_gray = cv2.imread("gray_sketch.png", 0)

line_mask = sketch_gray < 127
# Create white background, then overlay original colors on lines
color_sketch = np.full_like(original_img, 255)
color_sketch[line_mask] = original_img[line_mask]

cv2.imwrite("natural_color_sketch.png", color_sketch)

This method gives you a more realistic colored sketch where lines take on the hues from the original scene.

内容的提问来源于stack exchange，提问作者Rashmi