PythonProjects/Depth-Buffer-3D.py at master · ninjafrostpn/PythonProjects · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
import cv2
from math import radians, cos, sin
import numpy as np
from random import randrange as rand

w = 500
h = 500
size = (w, h)
allpoints = np.float32([[x, y, 0] for x in range(w) for y in range(h)]).reshape(-1, 1, 2)

allcolours = set()
palette = []
maxdepth = 400

cv2.namedWindow("")

# https://stackoverflow.com/a/46048098
# x is to the right, y is down, z is into the screen
obj_points = np.float32([0, 0, 10,
                         w, 0, 10,
                         w, h, 10,
                         0, h, 10,
                         w/2, h/2, 0]).reshape(1, -1, 3)
img_points = np.float32([100, 100,
                         w - 100, 100,
                         w - 100, h - 100,
                         100, h - 100,
                         w/2, h/2]).reshape(1, -1, 2)

camera_matrix = np.zeros((3, 3))
camera_matrix[0, 0] = 100  # F_y
camera_matrix[1, 1] = 100  # F_x
camera_matrix[2, 2] = 1.0
camera_matrix[0, 2] = h/2  # C_y
camera_matrix[1, 2] = w/2  # C_x
dist_coefs = np.zeros(4, dtype="float32")
retval, produced_camera_matrix, produced_dist_coefs, rvec, tvec = cv2.calibrateCamera(obj_points, img_points,
                                                                                      size, camera_matrix,
                                                                                      dist_coefs,
                                                                                      flags=cv2.CALIB_USE_INTRINSIC_GUESS)


def transformpoints(*points):
    point3D = np.float32(points).reshape(1, -1, 3)
    point2D, _ = cv2.projectPoints(point3D, np.float32(rvec), np.float32(tvec),
                                   produced_camera_matrix, produced_dist_coefs)
    return point2D[:, 0]


class Image:
    def __init__(self, image, alphamask=True):
        self.alphamask = alphamask
        self.colours = [tuple(colour) for colour in np.unique(image.reshape(-1, 3), axis=0)]
        self.colourmask = np.zeros(image.shape[:2], dtype='uint8')
        for colour in self.colours:
            print(colour)
            if colour not in allcolours:
                allcolours.add(colour)
                palette.append(colour)
            colourpos = (image == colour)[:, :].any(axis=2)
            colourindex = palette.index(colour)
            self.colourmask[colourpos] = colourindex
        self.w, self.h = self.colourmask.shape[:2]
        self.corners2D = np.float32([[0, 0],
                                    [self.w, 0],
                                    [self.w, self.h],
                                    [0, self.h]]).reshape(4, 1, 2)

    def render(self, y, x, z):
        points3D = np.array([[x, y, z],
                             [x + self.w, y, z],
                             [x + self.w, y + self.h, z],
                             [x, y + self.h, z]])
        points2D = transformpoints(points3D)
        xon = not (np.all(0 > points2D[:, 0]) or np.all(points2D[:, 0] > w))
        yon = not (np.all(0 > points2D[:, 1]) or np.all(points2D[:, 1] > h))
        if xon and yon:
            M1 = cv2.getPerspectiveTransform(self.corners2D, points2D)
            talpha = cv2.warpPerspective(np.float32(self.alphamask), M1, size[::-1],
                                         borderMode=cv2.BORDER_TRANSPARENT) > 0
            tcolour = cv2.warpPerspective(np.array(palette)[self.colourmask], M1, size[::-1],
                                          borderMode=cv2.BORDER_TRANSPARENT)

            M2 = cv2.getPerspectiveTransform(points2D, self.corners2D)
            M3, _ = cv2.findHomography(self.corners2D, points3D)
            tdepth = cv2.perspectiveTransform(cv2.perspectiveTransform(allpoints, M2), M3).reshape(*size, -1)[:, :, 2]

            drawpixels = (screendepth > tdepth) & talpha

            screendepth[drawpixels] = tdepth[drawpixels]
            screen[drawpixels] = tcolour[drawpixels]


imagealpha = np.ones((40, 40), dtype='bool')
imagealpha[10:21, 10:21] = 0
squareno = 16
squares = []
for i in range(squareno):
    imagecolour = np.zeros((40, 40, 3), dtype='uint8')
    imagecolour[:, :] = (0, rand(256), rand(256))
    squares.append(Image(imagecolour, imagealpha))

imagealpha = np.ones((80, h - 50), dtype='bool')
imagecolour = np.zeros((80, h - 50, 3), dtype='uint8')
imagecolour[:, :] = (rand(256), 0, rand(128))
tower = Image(imagecolour, imagealpha)

cycles = 0
while True:
    screendepth = np.zeros((w, h), dtype='float64')
    screendepth[:, :] = maxdepth
    screen = np.zeros((w, h, 3), dtype='uint8')
    for i, I in enumerate(squares):
        I.render((w/2 * (cos(radians(cycles + (i * (360/squareno)))) + 1)) - 20,
                 h/2,
                 (w/2 * (sin(radians(cycles + (i * (360/squareno)))) + 1)) - 20)
    tower.render(w/2 - 40, 0, 100)
    cv2.imshow("", cv2.cvtColor(np.rot90(np.fliplr(screen)), cv2.COLOR_RGB2BGR))
    cycles += 1
    key = cv2.waitKey(1)
    if key == 27:
        break
print("Nooooo...")