Shulou(Shulou.com)06/02 Report--

Today, I will talk to you about how to use the Python NumPy library to draw gradient patterns, which may not be well understood by many people. In order to make you understand better, the editor has summarized the following content for you. I hope you can get something according to this article.

Can NumPy also draw pictures? That's for sure! NumPy can not only draw, but also draw better and faster! For example, the following picture can be drawn in only 10 lines of code. If you can understand these 10 lines of code, it means knocking on the door of NumPy. Please open your Python IDLE and follow in my footsteps to experience the fun of interactive programming to see how to draw with NumPy and what kind of pictures can be drawn with NumPy.

1. Import module

Only import NumPy to complete the painting process, PIL's Image module is only used to display or save the painting results. It would be much easier and richer to deal with colors if you could invite Matplotlib's ColorMap to help, but that doesn't mean ColorMap is necessary.

> import numpy as np > from PIL import Image > from matplotlib import cm as mplcm2. Basic painting process

With the help of the Image.fromarray () function, you can convert the array generated by NumPy into a PIL object. The show () method of the PIL object can display the image directly, and the save () method can save the image as a file. In this series of operations, there is a very key knowledge point: the type of NumPy array must be a single-byte unsigned integer, that is, np.uint8 or np.ubyte type. The following code uses NumPy's random sub-module random to generate a two-dimensional array of 100 rows and 300 columns, which is converted into a random grayscale image with a width of 300 pixels and a height of 100 pixels and is displayed directly.

> im = np.random.randint (0,255, (100300), dtype=np.uint8) > im = Image.fromarray (im) > im.show () # or im.save (rnsd:\ gray_300_100.jpg') is saved as a file

3. Generate a random color image

In the above code, if the array generated by random contains three channels, you will get a random image in color.

> im = np.random.randint (0,255, (100,255,300), dtype=np.uint8) > Image.fromarray (im, mode='RGB') .show ()

4. Generate a gradient image

The np.linspace () function is similar to Python's range () function, which returns the isometric sequence of floating-point numbers. After np.tile () repetition, it generates a two-dimensional array of RGB channels, then merges it into a three-dimensional array with np.dstack (), and finally outputs a gradient image.

> r = np.tile (np.linspace (192jue 255,300, dtype=np.uint8), (600jue 1)) .T > > g = np.tile (np.linspace (192pyr255,600, dtype=np.uint8), (300jue 1)) > > b = np.ones ((300600), dtype=np.uint8) * 224 > im = np.dstack ((rjingjue b)) > > Image.fromarray (im, mode='RGB'). Show ()

5. Draw a curve on a gradient background

After positioning the specific rows and rows in the image array, and then changing their color, you can get the desired results.

> r = np.tile (np.linspace (192Yu255,300, dtype=np.uint8)) .T > g = np.tile (np.linspace (192Yu255,600, dtype=np.uint8), (300Yu1)) > > b = np.ones ((300600), dtype=np.uint8) * 224 > im = np.dstack ((rjingjinb)) > > x = np.arange (600) > > y = np.sin (0, 2*np.pi) For i in range (0150,6): im [y [:-I], [:-I]] = np.array ([255P0255]) > Image.fromarray (im, mode='RGB'). Show ()

6. Use color mapping (ColorMap)

Color mapping (ColorMap) is an indispensable concept of data visualization. It is after color mapping that boring data becomes colorful and pleasing to the eye. The cm sub-module of Matplotlib provides a total of 82 color mapping tables in 7 categories, each of which is appended with "_ r" to obtain an inverted version of the mapping table.

The following is the color band diagram of the jet color mapping table in the exclusive custom class and the Paired color mapping table in the segmented ladder class.

The cm submodule of Matplotlib is also very easy to use. The following code helps you understand the mechanism of color mapping (ColorMap) and familiarize yourself with the use of cm objects.

> cm1 = mplcm.get_cmap ('jet') # jet is the ColorMap of the exclusive custom class > cm1.N # jet has 256 colors (0.0,0.0,0.5,1.0) > > cm1 (128) # returns the color with serial number 128 (0.49019604313725,1.0,0.47758570524984,1.0) > > cm1 (05,05,0.05,0.25,0.0,0.05,0.25,0.05,0.01,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 Cm2 = mplcm.get_cmap ('Paired') # Paired is the segmented ladder ColorMap > cm2.N # Paired has 12 colors 12 > cm2 (0) # returns the color with serial number 0 (0.6509803921568628,0.807843137254902,0.89019607843137251.0) > > cm2 (11) # returns the color with serial number 11 (0.6941176470588235,0.349019604313724,0.1568627450980392,1.0) 7. Show the charm of NumPy

For an image (if the image is 9 pixels wide and 7 pixels high), you can easily get a two-dimensional array of the line number of each pixel (represented by I) and a two-dimensional array of the column number of each pixel (represented by j).

> > w, h = 9, 7 > I = np.repeat (np.arange (h), w) .reshape (h, w) > > j = np.tile (np.arange (w), (np.arange 1)) > iarray ([0,0,0,0,0,0,0,0,0,0], [1,1,1,1,1], [2,2,2,2,2,2], [3,3] 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 6, 6]) > > jarray ([0, 1, 2, 3, 4, 5, 6, 7, 8] 7, 8], [0,1,2,3,4,5,6,7,8], [0,1,2,3,4,5,6,7,8], [0,1,2,3,4,5,6,7,8], [0,1,2,3,4,5,6,7,8])

After a little transformation, the coordinates of each pixel in the plane Cartesian coordinate system with the center point of the image as the origin are obtained.

> I = I-h 2 > j = j-w Unix 2 > iarray ([[- 3,-3,-3,-3,-3,-3,-3,-3,-3,-3,-2,-2,-2,-2,-2,-2], [- 1,-1,-1,-1,-1,-1,-1] -1,-1], [0,0,0,0,0,0,0,0,0,0,0,0,0,0], [1,1,1,1], [2, 2, 2, 2, 2, 2], [3, 3, 3, 3, 3, 3] 3, 3]) > jarray ([- 4,-3,-2,-1, 0, 1, 2, 3, 4], [- 4,-3,-2,-1, 2, 3, 4], [- 4,-3,-2,-1, 2, 3, 4] -1, 0, 1, 2, 3, 4], [- 4,-3,-2,-1, 0, 1, 2, 3, 4], [- 4,-3,-2,-1, 2, 3, 4])

Naturally, it is also easy to calculate an array of distances (expressed in d) from the center of the image for each pixel. The following code uses the np.hypot () function to calculate the distance. It's no problem if you sum the square first and then square it. It's just not cool.

> > d = np.hypot (I, j) > darray ([[5., 4.24264069, 3.60555128, 3.16227766, 3., 3.16227766, 3.60555128, 4.24264069, 5.], [4.47213595, 3.60555128, 2.82842712, 2.23606798, 2., 2.23606798,2.82842712,3.6055128,4.47213595], [4.12310563, 3.16227766,2.23606798] 1.41421356, 1., 1.41421356, 2.23606798, 3.16227766, 4.12310563], [4., 3., 2., 1., 0. , 1., 2., 3., 4.], [4.12310563, 3.16227766, 2.23606798, 1.41421356, 1., 1.41421356, 2.23606798, 3.16227766, 4.12310563], [4.47213595, 3.60555128, 2.82842712, 2.23606798, 2., 2.23606798, 2.82842712, 3.60555128, 4.47213595] [5., 4.24264069, 3.60555128, 3.16227766, 3., 3.16227766, 3.60555128, 4.24264069, 5.])

Imagine what the image would look like if you wanted to map different distances to different colors using the jet color mapping table. What will the image look like if you select a specific area of the image, such as all pixels whose column number is less than 10 times the line number, and map the distance of each point in the selected area to different colors using the Paired color mapping table? This is done in 10 lines of code below.

> > def draw_picture (w, h, cm1='jet', cm2='Paired'): cm1, cm2= mplcm.get_cmap (cm1), mplcm.get_cmap (cm2) colormap1, colormap2 = np.array ([cm1 (k) for k in range (cm1.N)]), np.array ([cm2 (k) for k in range (cm2.N)]) I, j = np.repeat (np.arange (h), w) .reshape (np.arange (h), w)-hComp2 Np.tile (np.arange (w), (hmae1))-w-cm1='jet', cm2='Paired' 2d = np.hypot (I, j) e = d [(j*j/10) > draw_picture (1,900, cm1='jet', cm2='Paired') Do you know more about how to use the Python NumPy library to draw gradient patterns? If you want to know more knowledge or related content, please follow the industry information channel, thank you for your support.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.