假设我有80个(或n
)极坐标,它们在圆形区域内均匀分布。我希望每个极坐标都有一种独特的颜色。
如果你想象一个像这样的色轮(如果你愿意,它可能是一个不同的转换),我喜欢它给出极坐标的一种颜色。
起初我没有使用实际的极坐标,只是通过一些甚至步幅缩放其中一个通道,如RGB (255, i * stride, 255)
。但是现在我喜欢不同频谱的不同颜色(或者至少有一种以上的色调)。
我想过只使用一个色轮图像然后对其进行采样,但这似乎有点弱。是不是有一个公式我可以用来将极坐标转换为某个假设/生成的RGB,HSV或CMYK空间?
我在Python 3工作,但我最感兴趣的是公式/算法。我没有使用任何特定的绘图API。
答案 0 :(得分:0)
您可以使用从HSV或HSL到RGB的转换,许多包,例如Colour(Numpy Vectorised)或python-colormath(Vanilla Python)都有实现:
从颜色,假设您有Numpy以及tsplit和tstack定义:
def RGB_to_HSV(RGB):
"""
Converts from *RGB* colourspace to *HSV* colourspace.
Parameters
----------
RGB : array_like
*RGB* colourspace array.
Returns
-------
ndarray
*HSV* array.
Notes
-----
- Input *RGB* colourspace array is in domain [0, 1].
- Output *HSV* colourspace array is in range [0, 1].
References
----------
- :cite:`EasyRGBj`
- :cite:`Smith1978b`
- :cite:`Wikipediacg`
Examples
--------
>>> RGB = np.array([0.49019608, 0.98039216, 0.25098039])
>>> RGB_to_HSV(RGB) # doctest: +ELLIPSIS
array([ 0.2786738..., 0.744 , 0.98039216])
"""
maximum = np.amax(RGB, -1)
delta = np.ptp(RGB, -1)
V = maximum
R, G, B = tsplit(RGB)
S = np.asarray(delta / maximum)
S[np.asarray(delta == 0)] = 0
delta_R = (((maximum - R) / 6) + (delta / 2)) / delta
delta_G = (((maximum - G) / 6) + (delta / 2)) / delta
delta_B = (((maximum - B) / 6) + (delta / 2)) / delta
H = delta_B - delta_G
H = np.where(G == maximum, (1 / 3) + delta_R - delta_B, H)
H = np.where(B == maximum, (2 / 3) + delta_G - delta_R, H)
H[np.asarray(H < 0)] += 1
H[np.asarray(H > 1)] -= 1
H[np.asarray(delta == 0)] = 0
HSV = tstack((H, S, V))
return HSV
def HSV_to_RGB(HSV):
"""
Converts from *HSV* colourspace to *RGB* colourspace.
Parameters
----------
HSV : array_like
*HSV* colourspace array.
Returns
-------
ndarray
*RGB* colourspace array.
Notes
-----
- Input *HSV* colourspace array is in domain [0, 1].
- Output *RGB* colourspace array is in range [0, 1].
References
----------
- :cite:`EasyRGBn`
- :cite:`Smith1978b`
- :cite:`Wikipediacg`
Examples
--------
>>> HSV = np.array([0.27867384, 0.74400000, 0.98039216])
>>> HSV_to_RGB(HSV) # doctest: +ELLIPSIS
array([ 0.4901960..., 0.9803921..., 0.2509803...])
"""
H, S, V = tsplit(HSV)
h = np.asarray(H * 6)
h[np.asarray(h == 6)] = 0
i = np.floor(h)
j = V * (1 - S)
k = V * (1 - S * (h - i))
l = V * (1 - S * (1 - (h - i))) # noqa
i = tstack((i, i, i)).astype(np.uint8)
RGB = np.choose(
i, [
tstack((V, l, j)),
tstack((k, V, j)),
tstack((j, V, l)),
tstack((j, k, V)),
tstack((l, j, V)),
tstack((V, j, k)),
],
mode='clip')
return RGB
def RGB_to_HSL(RGB):
"""
Converts from *RGB* colourspace to *HSL* colourspace.
Parameters
----------
RGB : array_like
*RGB* colourspace array.
Returns
-------
ndarray
*HSL* array.
Notes
-----
- Input *RGB* colourspace array is in domain [0, 1].
- Output *HSL* colourspace array is in range [0, 1].
References
----------
- :cite:`EasyRGBl`
- :cite:`Smith1978b`
- :cite:`Wikipediacg`
Examples
--------
>>> RGB = np.array([0.49019608, 0.98039216, 0.25098039])
>>> RGB_to_HSL(RGB) # doctest: +ELLIPSIS
array([ 0.2786738..., 0.9489796..., 0.6156862...])
"""
minimum = np.amin(RGB, -1)
maximum = np.amax(RGB, -1)
delta = np.ptp(RGB, -1)
R, G, B = tsplit(RGB)
L = (maximum + minimum) / 2
S = np.where(L < 0.5, delta / (maximum + minimum),
delta / (2 - maximum - minimum))
S[np.asarray(delta == 0)] = 0
delta_R = (((maximum - R) / 6) + (delta / 2)) / delta
delta_G = (((maximum - G) / 6) + (delta / 2)) / delta
delta_B = (((maximum - B) / 6) + (delta / 2)) / delta
H = delta_B - delta_G
H = np.where(G == maximum, (1 / 3) + delta_R - delta_B, H)
H = np.where(B == maximum, (2 / 3) + delta_G - delta_R, H)
H[np.asarray(H < 0)] += 1
H[np.asarray(H > 1)] -= 1
H[np.asarray(delta == 0)] = 0
HSL = tstack((H, S, L))
return HSL