[toc]
初衷
曼德布洛特集合是在李善友老师的混沌大学公开课上听到的,在极客学院使用TensorFlow也可以实现该集合,将其可视化时,于是尝试模拟该模型。
参考文档
Numpy、matplotlib实现二维数据到图像的转换,添加colormap,无边距显示:https://blog.csdn.net/u010105243/article/details/76695400
极客学院源码曼德布洛特集合源码(较老):http://wiki.jikexueyuan.com/project/tensorflow-zh/tutorials/mandelbrot.html
源码实现曼德勃罗集合
"""
本案例来展示偏微分方程
使用tensorflow来实现该模型
"""
import tensorflow as tf
import numpy as np
import PIL.Image
import PIL.ImageDraw
import matplotlib.pyplot as plt
def DisplayFractal(a, fmt='jpeg'):
"""显示迭代计算出的彩色分形图像。"""
a_cyclic = (6.28*a/20.0).reshape(list(a.shape)+[1])
img = np.concatenate([10+20*np.cos(a_cyclic),
30+50*np.sin(a_cyclic),
155-80*np.cos(a_cyclic)], 2)
img[a==a.max()] = 0
a = img
a = np.uint8(np.clip(a, 0, 255))
img1 = PIL.Image.fromarray(a)
plt.imsave("image_tf.png", img1)
plt.show()
sess = tf.InteractiveSession()
# 使用NumPy创建一个在[-2,2]x[-2,2]范围内的2维复数数组
Y, X = np.mgrid[-1.3:1.3:0.005, -2:1:0.005]
Z = X+1j*Y
xs = tf.constant(Z.astype("complex64"))
zs = tf.Variable(xs)
ns = tf.Variable(tf.zeros_like(xs, "float32"))
tf.global_variables_initializer().run()
# 计算一个新值z: z^2 + x
zs_ = zs*zs + xs
# 这个新值会发散吗?
not_diverged = tf.abs(zs_) < 4
# 更新zs并且迭代计算。
#
# 说明:在这些值发散之后,我们仍然在计算zs,这个计算消耗特别大!
# 如果稍微简单点,这里有更好的方法来处理。
#
step = tf.group(
zs.assign(zs_),
ns.assign_add(tf.cast(not_diverged, "float32"))
)
for i in range(200):
step.run()
DisplayFractal(ns.eval())
这里对wiki上的源码做了三处改动
1. import matplotlib.pyplot as plt
使用plt绘制二维数组图
2. img1 = PIL.Image.fromarray(a)
plt.imsave("image_tf.png", img1)
用plt绘图,而不是IPython.display
3. # 这个新值会发散吗?
not_diverged = tf.abs(zs_) < 4
原函数tf.complex_abs已经在新版tf中弃用。
结果如下,在当前目录下生成一张图片image_tf.png:
后记
还有其他绘制曼德勃罗集合的方式,这里就不展开了。
"""
本案例用来展示曼德布洛特迭代模型.
使用纯粹的python 算法实现该模型, machine_10 使用tensorflow来实现该模型
绘图使用PIL
"""
import time
from PIL import Image, ImageDraw
g_size = (400, 300) # 图形最终尺寸
g_max_iteration = 256 # 最大迭代次数
g_bailout = 4 # 最大域
g_zoom = 2.5 / g_size[0] # 缩放参数
g_offset = (-g_size[0] * 0.25, 0) # 偏移量
g_HSL = (210, 80, 50) # HSL色彩基调
def draw(antialias = True):
zi = 2 if antialias else 1 # antialias: 抗锯齿 size = [i * zi
size = [i * zi for i in g_size]
zoom = g_zoom / zi
offset = [i * zi for i in g_offset]
bailout = g_bailout * zi
img = Image.new("RGB", size, 0xffffff)
dr = ImageDraw.Draw(img)
print("painting Mandelbrot Set..")
for xy, color in getPoints(size, offset, zoom):
dr.point(xy, fill = color)
print("100%n")
del dr
if antialias:
img = img.resize(g_size, Image.ANTIALIAS)
img.show()
img.save("mandelbrot_set_%dx%d.png" % g_size)
def getPoints(size, offset, zoom, ti = 0, tstep = 1):
"生成需要绘制的点的坐标及颜色"
def getRepeats(c):
z = c
repeats = 0
while abs(z) < g_bailout and repeats < g_max_iteration:
z = z * z + c
repeats += 1
return repeats
def getColor(r):
color = "hsl(0, 0%, 0%)"
if r < g_max_iteration:
v = 1.0 * r / g_max_iteration
h = ch * (1 - v)
s = cs
l = cl * (1 + v)
color = "hsl(%d, %d%%, %d%%)" % (h, s, l)
return color
xs, ys = size
xw, yh = xs / 2, ys / 2
xo, yo = offset
ch, cs, cl = g_HSL
progress = 0
for iy in range(ys):
p = iy * 100 / ys
if iy % 10 == 0 and p != progress:
print ("%d%%..." % p) # 显示进度
progress = p
for ix in range(ti, xs, tstep):
x = (ix - xw + xo) * zoom
y = (iy - yh + yo) * zoom
c = complex(x, y)
r = getRepeats(c)
yield (ix, iy), getColor(r)
def main():
t0 = time.time()
draw()
t = time.time() - t0
print("%dm%.3fs" % (t / 60, t % 60))
if __name__ == "__main__":
main()