-
Notifications
You must be signed in to change notification settings - Fork 1
/
edge.py
198 lines (175 loc) · 8.03 KB
/
edge.py
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
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
import cv2
import numpy as np
import matplotlib.pyplot as plt
import math
import time
# from skimage.morphology import skeletonize
# 使用霍夫直线变换做直线检测，前提条件：边缘检测已经完成
# 统计概率霍夫线变换
def offside_dectet(image, direction, ofplayers, dfplayer):
debug = 0
img_origin = image.copy()
shrink1 = 2
shrink2 = 4
if direction in ['left','up']:
dfplayer_x = dfplayer[0]
dfplayer_y = dfplayer[1]
else:
dfplayer_x = dfplayer[0]+dfplayer[2]
dfplayer_y = dfplayer[1]+dfplayer[3]
image = cv2.resize(image,(math.ceil(image.shape[1]/shrink1),math.ceil(image.shape[0]/shrink1)))
has_offside = []
th = 20 # 边缘检测后大于th的才算边界
gray = cv2.cvtColor(image, cv2.COLOR_BGRA2GRAY)
gray_origin = cv2.cvtColor(img_origin, cv2.COLOR_BGRA2GRAY)
gray = cv2.GaussianBlur(gray, (5, 5), 0)
x = cv2.Sobel(gray, cv2.CV_16S, 1, 0) # x方向梯度
y = cv2.Sobel(gray, cv2.CV_16S, 0, 1) # y方向梯度
absX = cv2.convertScaleAbs(x) # 转回uint8
absY = cv2.convertScaleAbs(y)
edges = cv2.addWeighted(absX, 0.5, absY, 0.5, 0) # 各0.5的权重将两个梯度叠加
dst, edges = cv2.threshold(edges, th, 255, cv2.THRESH_BINARY) # 大于th的赋值255（白色）
edges = cv2.resize(edges, (math.ceil(image.shape[1] / shrink2), math.ceil(image.shape[0] / shrink2)))
if debug == 1:
cv2.imshow('edge', edges)
# 函数将通过步长为1的半径和步长为π/180的角来搜索所有可能的直线
lines = cv2.HoughLinesP(edges, 1, np.pi / 180, 50, minLineLength=min(edges.shape[0], edges.shape[1])/2,
maxLineGap=math.ceil(40/shrink2))
# print(lines)
angle = [] # 备选线的角度
b = [] # 备选线 y=kx+b的b
if lines is None:
has_line = 0
else:
has_line = 1
for line in lines:
x1, y1, x2, y2 = line[0]
angle_per = math.atan((y2 - y1) / (x2 - x1)) # 角度
if angle_per < -np.pi / 4: # 将角度换到-pi/4 ~ 3pi/4
angle_per = angle_per + np.pi
angle.append(angle_per)
# b.append(x1 * (y2 - y1) / (x2 - x1) - y1)
b.append(shrink1*shrink2*(x2*y1-x1*y2) / (x2 - x1) )
if debug == 1:
cv2.line(img_origin, (x1*shrink1*shrink2, y1*shrink1*shrink2), (x2*shrink1*shrink2, y2*shrink1*shrink2), (0, 0, 255), 1) # 画线
angle = np.array(angle)
b = np.array(b)
threshold = 0.3
if direction=='up' or direction=='down':
angle_delete_vertical = angle[(angle<threshold) & (angle>-threshold)]
b_delete_vertical = b[(angle<threshold) & (angle>-threshold)]
elif direction == 'left' or direction=='right':
angle_delete_vertical = angle[(angle < np.pi/2 + threshold) & (angle > np.pi/2 - threshold)]
b_delete_vertical = b[(angle < np.pi/2 + threshold) & (angle > np.pi/2 - threshold)]
# print(angle_delete_vertical)
# angle_ave = np.median(angle_delete_vertical) # 角度平均值
angle_ave = np.median(angle_delete_vertical) # 角度中位数
angle_diff = angle_delete_vertical - angle_ave # 与平均值的差
b = b_delete_vertical[abs(angle_diff) < 0.08] # 去除离群点
angle = angle_delete_vertical[abs(angle_diff) < 0.08] # 去除离群点
# print(angle)
if len(angle) == 0:
has_line = 0
else:
k_unsort = np.tan(angle) # 角度对应的k
b = np.array(b)
dis = abs(k_unsort * dfplayer_x - dfplayer_y + b) / np.sqrt(1 + k_unsort * k_unsort) # 防守球员到线的距离
dis = list(dis)
angle_final = angle[dis.index(min(dis))] # 选择离防守球员最近的线
if abs(angle_final) < 0.001: # 处理奇异情况
if angle_final < 0:
angle_final = -0.001
else:
angle_final = 0.001
elif abs(angle_final) > 1.56 and abs(angle_final) < 1.58:
angle_final = 1.56 * angle_final / abs(angle_final)
k = np.tan(angle_final) # 最终的k
# k = 0.10422
for ofplayer in ofplayers:
if direction in ['left', 'up']:
ofplayer_x = ofplayer[0]
ofplayer_y = ofplayer[1]
else:
ofplayer_x = ofplayer[0] + ofplayer[2]
ofplayer_y = ofplayer[1] + ofplayer[3]
# # 画出越位线
# y1_draw = int(dfplayer_y - k * dfplayer_x)
# y2_draw = int(k * gray_origin.shape[1] - k * dfplayer_x + dfplayer_y)
# if debug==1:
# cv2.line(img_origin, (0, y1_draw), (gray_origin.shape[1], y2_draw), (0, 255, 0), 1)
# # 画出防守球员和进攻球员
# cv2.circle(img_origin, (dfplayer_x, dfplayer_y), 5, (255, 0, 0))
# cv2.circle(img_origin, (ofplayer_x, ofplayer_y), 5, (255, 0, 0))
# 越位判罚
line_x = ofplayer_y - (dfplayer_y - k * dfplayer_x) / k
line_y = k * ofplayer_x - k * dfplayer_x + dfplayer_y
if direction == 'left':
if line_x > ofplayer_x:
has_offside.append(1)
else:
has_offside.append(0)
elif direction == 'right':
if line_x < ofplayer_x:
has_offside.append(1)
else:
has_offside.append(0)
elif direction == 'up':
if line_y > ofplayer_y:
has_offside.append(1)
else:
has_offside.append(0)
elif direction == 'down':
if line_y < ofplayer_y:
has_offside.append(1)
else:
has_offside.append(0)
if debug == 1:
cv2.imshow("line_detect_possible_demo", img_origin)
return k,has_line, has_offside
def draw_offside_line(img_origin,direction,dfplayer,k):
debug = 1
if direction in ['left','up']:
dfplayer_x = dfplayer[0]
dfplayer_y = dfplayer[1]
else:
dfplayer_x = dfplayer[0]+dfplayer[2]
dfplayer_y = dfplayer[1]+dfplayer[3]
# 画出越位线
y1_draw = int(dfplayer_y - k * dfplayer_x)
y2_draw = int(k * img_origin.shape[1] - k * dfplayer_x + dfplayer_y)
if debug == 1:
cv2.line(img_origin, (0, y1_draw), (img_origin.shape[1], y2_draw), (242, 232, 22), 2)
# 画出防守球员和进攻球员
cv2.circle(img_origin, (dfplayer_x, dfplayer_y), 5, (255, 0, 0))
# cv2.circle(img_origin, (ofplayer_x, ofplayer_y), 5, (255, 0, 0))
return img_origin
if __name__ == "__main__":
cap = cv2.VideoCapture('/home/jiangcx/桌面/足球视频/offside2.mp4')
while (cap.isOpened()):
print('-----frame#-----')
ret, img = cap.read()
# img = cv2.imread('edge.png')
img = cv2.resize(img,(1920,1080))
# cv2.imshow('original', img)
# img = cv2.imread('edge16.png')
# （图像，向哪个方向进攻(left & right)，进攻球员x，进攻球员y，防守球员x，防守球员y
start_time = time.time()
ofplayers = np.array([[10,20]])
deplayer = np.array([100, 200])
k,has_line, has_offsides = offside_dectet(img, 'up', ofplayers, deplayer)
draw_offside_line(img, "up", deplayer, k)
time1 = time.time()
print('time1', time1 - start_time)
# time2 = time.time()
# print('time2', time2- time1)
for has_offside in has_offsides:
if has_line == 1:
print('has_line')
if has_offside == 1:
print('越位')
else:
print('不越位')
else:
print("no_line")
cv2.waitKey(-1)
cv2.destroyAllWindows()