Creating a Rotating Earth GIF with Python, Pillow, NumPy, and imageio

The article explains how to create a simple yet common rotating Earth animation using a 360° latitude‑longitude texture and Python.

First, obtain a 360° equirectangular map where the height is half the width, representing the Earth's surface from pole to pole.

The rotation effect is achieved by projecting each pixel of the texture onto a circle, then stitching the resulting frames into a GIF.

The process consists of five steps: (1) select all points inside a circle, (2) compute the latitude and longitude for each point, (3) map those coordinates back to x‑y positions on the texture, (4) write the corresponding pixel colors to a new image, and (5) assemble the frames into a GIF.

Below is the code for selecting points inside the circle:

img = Image.new('RGBA', (300,300), 'black')
w = img.size[0]
h = img.size[1]
pxList = []
pyList = []
for i in range(w):
    for j in range(h):
        r = math.sqrt((i-w/2)**2 + (j-h/2)**2)
        if r < 150:
            pxList.append(i)
            pyList.append(j)

The function calcSphereXY2XYZ converts planar coordinates to spherical latitude and longitude using NumPy for speed:

def calcSphereXY2XYZ(px, py, maxHeight, longOffset):
    v0x = np.array(px)
    v0y = np.array(py)
    v03 = np.subtract(v0x, maxHeight)
    v04 = np.subtract(v0y, maxHeight)
    v1x = np.true_divide(v03, maxHeight)
    v1y = np.true_divide(v04, maxHeight)
    v07 = np.power(v1x,2)
    v08 = np.power(v1y,2)
    v09 = np.add(v07, v08)
    v0a = np.subtract(1, v09)
    v1z = np.power(v0a, 1/2)  # z
    v1lat = np.multiply(v1y, math.pi/2)  # lat
    v0lon = np.arctan2(v1z, -v1x)
    v1lon = np.add(v0lon, longOffset)  # long
    v2lon = np.fmod(v1lon, math.pi*2)  # long
    return v2lon, v1lat

The function calcShpereLatLong2XY maps latitude‑longitude to image coordinates:

def calcShpereLatLong2XY(vlon, vlat, width, height):
    v3x0 = np.multiply(vlon, width/2/math.pi)
    v3y0 = np.multiply(vlat, height/math.pi)
    v3y1 = np.add(v3y0, height/2)
    v3x2 = v3x0.astype(np.integer)
    v3y2 = v3y1.astype(np.integer)
    return v3x2, v3y2

After loading the texture and converting it to a NumPy array, the pixel colors are transferred to the circular image:

color = arrayBack[npy, npx]
for i in range(len(pxList)):
    x = pxList[i]
    y = pyList[i]
    cc = color[i]
    cc = tuple(cc)
    img.putpixel((x, y), cc)

Finally, the frames are saved and combined into an animated GIF using the imageio library:

frames = []
for i in range(0, 360, 10):
    a = -i * math.pi / 180
    img = getPic(a)
    filename = 'temp%03d.png' % i
    img.save(filename)
    im = imageio.imread(filename)
    frames.append(im)
imageio.mimsave('earth.gif', frames, 'GIF', duration=0.25)

The article concludes with the full source code that integrates all the steps into a single script.