利用Python將圖片批量轉(zhuǎn)化成素描圖的過(guò)程記錄

更新時(shí)間：2021年08月11日 11:00:18 作者：guihunkun

萬(wàn)能的Python真的是除了不會(huì)生孩子,其他的還真不在話下,下面這篇文章主要給大家介紹了關(guān)于如何利用Python將圖片批量轉(zhuǎn)化成素描圖的相關(guān)資料,需要的朋友可以參考下

前言

正常圖片轉(zhuǎn)化成素描圖片無(wú)非對(duì)圖片像素的處理，矩陣變化而已。目前很多拍照修圖App都有這一功能，核心代碼不超30行。如下利用 Python 實(shí)現(xiàn)讀取一張圖片并將其轉(zhuǎn)化成素描圖片。至于批處理也簡(jiǎn)單，循環(huán)讀取文件夾里的圖片處理即可。具體代碼可以去我的 GitHub 下載。

程序

Method 1

def plot_sketch(origin_picture, out_picture) :
    a = np.asarray(Image.open(origin_picture).convert('L')).astype('float')
    depth = 10.  # (0-100)
    grad = np.gradient(a)  # 取圖像灰度的梯度值
    grad_x, grad_y = grad  # 分別取橫縱圖像梯度值
    grad_x = grad_x * depth / 100.
    grad_y = grad_y * depth / 100.
    A = np.sqrt(grad_x ** 2 + grad_y ** 2 + 1.0)
    uni_x = grad_x / A
    uni_y = grad_y / A
    uni_z = 1. / A

    vec_el = np.pi / 2.2  # 光源的俯視角度，弧度值
    vec_az = np.pi / 4.  # 光源的方位角度，弧度值
    dx = np.cos(vec_el) * np.cos(vec_az)  # 光源對(duì)x 軸的影響
    dy = np.cos(vec_el) * np.sin(vec_az)  # 光源對(duì)y 軸的影響
    dz = np.sin(vec_el)  # 光源對(duì)z 軸的影響

    b = 255 * (dx * uni_x + dy * uni_y + dz * uni_z)  # 光源歸一化
    b = b.clip(0, 255)

    im = Image.fromarray(b.astype('uint8'))  # 重構(gòu)圖像
    im.save(out_picture)
    print("轉(zhuǎn)換成功，請(qǐng)查看 ： ", out_picture)

Method 2

def plot_sketch2(origin_picture, out_picture, alpha=1.0):
    img = Image.open(origin_picture)
    blur = 20
    img1 = img.convert('L')  # 圖片轉(zhuǎn)換成灰色
    img2 = img1.copy()
    img2 = ImageOps.invert(img2)
    for i in range(blur):  # 模糊度
        img2 = img2.filter(ImageFilter.BLUR)
    width, height = img1.size
    for x in range(width):
        for y in range(height):
            a = img1.getpixel((x, y))
            b = img2.getpixel((x, y))
            img1.putpixel((x, y), min(int(a*255/(256-b*alpha)), 255))
    img1.save(out_picture)

完整代碼

from PIL import Image, ImageFilter, ImageOps
import numpy as np
import os


def plot_sketch(origin_picture, out_picture) :
    a = np.asarray(Image.open(origin_picture).convert('L')).astype('float')
    depth = 10.  # (0-100)
    grad = np.gradient(a)  # 取圖像灰度的梯度值
    grad_x, grad_y = grad  # 分別取橫縱圖像梯度值
    grad_x = grad_x * depth / 100.
    grad_y = grad_y * depth / 100.
    A = np.sqrt(grad_x ** 2 + grad_y ** 2 + 1.0)
    uni_x = grad_x / A
    uni_y = grad_y / A
    uni_z = 1. / A

    vec_el = np.pi / 2.2  # 光源的俯視角度，弧度值
    vec_az = np.pi / 4.  # 光源的方位角度，弧度值
    dx = np.cos(vec_el) * np.cos(vec_az)  # 光源對(duì)x 軸的影響
    dy = np.cos(vec_el) * np.sin(vec_az)  # 光源對(duì)y 軸的影響
    dz = np.sin(vec_el)  # 光源對(duì)z 軸的影響

    b = 255 * (dx * uni_x + dy * uni_y + dz * uni_z)  # 光源歸一化
    b = b.clip(0, 255)

    im = Image.fromarray(b.astype('uint8'))  # 重構(gòu)圖像
    im.save(out_picture)
    print("轉(zhuǎn)換成功，請(qǐng)查看 ： ", out_picture)


def plot_sketch2(origin_picture, out_picture, alpha=1.0):
    img = Image.open(origin_picture)
    blur = 20
    img1 = img.convert('L')  # 圖片轉(zhuǎn)換成灰色
    img2 = img1.copy()
    img2 = ImageOps.invert(img2)
    for i in range(blur):  # 模糊度
        img2 = img2.filter(ImageFilter.BLUR)
    width, height = img1.size
    for x in range(width):
        for y in range(height):
            a = img1.getpixel((x, y))
            b = img2.getpixel((x, y))
            img1.putpixel((x, y), min(int(a*255/(256-b*alpha)), 255))
    img1.save(out_picture)


if __name__ == '__main__':
    origin_picture = "pictures/5.jpg"
    out_picture = "sketchs/sketch.jpg"
    plot_sketch(origin_picture, out_picture)

    origin_path = "./pictures"
    out_path = "./sketchs"
    dirs = os.listdir(origin_path)
    for file in dirs:
        origin_picture = origin_path + "/" + file
        out_picture = out_path + "/" + "sketch_of_" + file
        plot_sketch2(origin_picture, out_picture)