问题：我使用三维数组创建三维模型，在使用rotate函数时，用了三组rotate函数，但是只有一组发挥了效果，请问这个是为什么呢？
我使用这个库来调用rotate函数
from scipy.ndimage import rotate
这里是部分代码，先这样的rotate函数我用了三次，但是只有一次是有效果的，请问如何解决呢？
cuboid = np.zeros ([20,30,50])
space = np.full ([100,100,100],255)
cuboid_x = random.randint(0,80)
cuboid_y = random.randint(0,70)
cuboid_z = random.randint(0,50)
space[cuboid_x: cuboid_x+20,cuboid_y: cuboid_y+30,cuboid_z: cuboid_z+50] = cuboid
angle_cuboid_x = random.randint(0, 359)
angle_cuboid_y = random.randint(0, 359)
angle_cuboid_z = random.randint(0, 359)
rotated_cuboid = rotate(cuboid, angle_cuboid_x, axes=(0, 1), reshape=False)
rotated_cuboid = rotate(rotated_cuboid, angle_cuboid_y, axes=(1, 2), reshape=False)
rotated_cuboid = rotate(rotated_cuboid, angle_cuboid_z, axes=(0, 2), reshape=False)
space[cuboid_x:cuboid_x+20, cuboid_y:cuboid_y+30 ,cuboid_z:cuboid_z+50] = rotated_cuboid
import numpy as np
# 旋转，axis为旋转轴，0,1,2分别代表x,y,z轴
# theta为旋转角度，单位已改为度，非弧度
# center为旋转中心，其为一维np数组[x,y,z]，默认值为图像中心点
def rotation(data, axis, theta, c = np.array([])):# c代表旋转点
theta = -np.pi * theta / 180
if c.size == 0:
c = np.array([np.floor((data.shape[0]-1)/2), np.floor((data.shape[1]-1)/2), np.floor((data.shape[1]-1)/2)])
s = data.shape
mean = np.mean(data)
# new_data = np.ones(s) * mean # 补均值
new_data = np.zeros(s) # 补零
# 绕x轴旋转
if axis == 0:
for i in range(0, s[0]):
for j in range(0, s[1]):
for k in range(0, s[2]):
x = i
y = (j-c[1])*np.cos(theta)-(k-c[2])*np.sin(theta)+c[1]
if(y < 0 or y > s[1]-1):
continue
z = (j-c[1])*np.sin(theta)+(k-c[2])*np.cos(theta)+c[2]
if(z < 0 or z > s[2]-1):
continue
y1 = np.floor(y).astype(int)
y2 = np.ceil(y).astype(int)
z1 = np.floor(z).astype(int)
z2 = np.ceil(z).astype(int)
dy = y - y1
dz = z - z1
new_data[i,j,k] = (data[x,y1,z1]*(1-dy)+data[x,y2,z1]*dy)*(1-dz) + (data[x,y1,z2]*(1-dy)+data[x,y2,z2]*dy)*dz
# 绕y轴旋转
elif axis == 1:
for i in range(0, s[0]):
for j in range(0, s[1]):
for k in range(0, s[2]):
y = j
x = (i-c[0])*np.cos(theta)-(k-c[2])*np.sin(theta)+c[0]
if(x < 0 or x > s[0]-1):
continue
z = (i-c[0])*np.sin(theta)+(k-c[2])*np.cos(theta)+c[2]
if(z < 0 or z > s[2]-1):
continue
x1 = np.floor(x).astype(int)
x2 = np.ceil(x).astype(int)
z1 = np.floor(z).astype(int)
z2 = np.ceil(z).astype(int)
dx = x - x1
dz = z - z1
new_data[i,j,k] = (data[x1,y,z1]*(1-dx)+data[x2,y,z1]*dx)*(1-dz) + (data[x1,y,z2]*(1-dx)+data[x2,y,z2]*dx)*dz
# 绕z轴旋转
else:
for i in range(0, s[0]):
for j in range(0, s[1]):
for k in range(0, s[2]):
z = k
x = (i-c[0])*np.cos(theta)-(j-c[1])*np.sin(theta)+c[0]
if(x < 0 or x > s[0]-1):
continue
y = (i-c[0])*np.sin(theta)+(j-c[1])*np.cos(theta)+c[1]
if(y < 0 or y > s[1]-1):
continue
x1 = np.floor(x).astype(int)
x2 = np.ceil(x).astype(int)
y1 = np.floor(y).astype(int)
y2 = np.ceil(y).astype(int)
dx = x - x1
dy = y - y1
new_data[i,j,k] = (data[x1,y1,z]*(1-dx)+data[x2,y1,z]*dx)*(1-dy) + (data[x1,y2,z]*(1-dx)+data[x2,y2,z]*dx)*dy
return new_data