-
Notifications
You must be signed in to change notification settings - Fork 0
/
turn_prediction.py
321 lines (246 loc) · 11.5 KB
/
turn_prediction.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
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
#!/usr/bin/python3
import numpy as np
# import matplotlib.pyplot as plt
from numpy.linalg import inv
import cv2 as cv
# GLobal params
num_windows = 10
margin=110
minpix=20
poly_points_1= np.array([[0,0], [0,720],[540,440],[740,440],[1280,720],[1280,0]])
poly_points_2= np.array([[0,652], [0,720],[1280,720],[1280,652]])
array_for_avg = []
array_for_avg_prediction = []
# Source points for homography
src = np.array([[605,440],[720,440], [1146, 651], [210, 690]], dtype="float32")
# Destination points for homography
dst = np.array([[0, 0], [250, 0], [250, 500], [0, 500]], dtype="float32")
#Homography
H_matrix = cv.getPerspectiveTransform(src, dst)
Hinv = inv(H_matrix)
def rolling_avg(arr):
global array_for_avg
array_for_avg.append(arr)
if len(array_for_avg) < 10:
array_for_avg.append(arr)
return arr
else:
array_for_avg.pop(0)
array_for_avg.append(arr)
return np.int32(np.mean(np.array(array_for_avg),axis=0))
def rolling_avg_prediction(arr):
global array_for_avg_prediction
array_for_avg_prediction.append(arr)
if len(array_for_avg_prediction) < 10:
array_for_avg_prediction.append(arr)
return arr
else:
array_for_avg_prediction.pop(0)
array_for_avg_prediction.append(arr)
return np.int32(np.mean(np.array(array_for_avg_prediction),axis=0))
def mask_poly(frame,poly_points):
img=frame.copy()
img = cv.fillPoly(img,[poly_points],(0,0,0))
# cv.namedWindow('post_mask',cv.WINDOW_NORMAL)
# cv.imshow('post_mask',img)
return img
def correct_dist(initial_img):
# Intrnsic camera matrix
k = [[1.15422732e+03, 0.00000000e+00, 6.71627794e+02], [0.00000000e+00, 1.14818221e+03, 3.86046312e+02],
[0.00000000e+00, 0.00000000e+00, 1.00000000e+00]]
k = np.array(k)
# Distortion Matrix
dist = [[-2.42565104e-01, -4.77893070e-02, -1.31388084e-03, -8.79107779e-05, 2.20573263e-02]]
dist = np.array(dist)
img_2 = cv.undistort(initial_img, k, dist, None, k)
return img_2
# Turn prediction for lanes
def turn_predict(image_center, right_lane_pos, left_lane_pos):
lane_center = left_lane_pos + (right_lane_pos - left_lane_pos)/2
if (lane_center - image_center < 0):
return ("Right Turn")
elif (lane_center - image_center < 8):
return ("Straight Road")
else:
return ("Left Turn")
# Process image and homography operations
def image_preprocessing(img):
# crop_img = img[420:720, 40:1280, :] # To get the region of interest
frame_post_mask = mask_poly(img,poly_points_1)
crop_img = mask_poly(frame_post_mask,poly_points_2)
# undist_img = correct_dist(crop_img)
hsl_img = cv.cvtColor(crop_img, cv.COLOR_BGR2HLS)
# To seperate out Yellow colored lanes
lower_mask_yellow = np.array([20, 120, 80], dtype='uint8')
upper_mask_yellow = np.array([45, 200, 255], dtype='uint8')
mask_yellow = cv.inRange(hsl_img, lower_mask_yellow, upper_mask_yellow)
yellow_detect = cv.bitwise_and(hsl_img, hsl_img, mask=mask_yellow).astype(np.uint8)
# To seperate out White colored lanes
lower_mask_white = np.array([0, 200, 0], dtype='uint8')
upper_mask_white = np.array([255, 255, 255], dtype='uint8')
mask_white = cv.inRange(hsl_img, lower_mask_white, upper_mask_white)
white_detect = cv.bitwise_and(hsl_img, hsl_img, mask=mask_white).astype(np.uint8)
# Combine both
lanes = cv.bitwise_or(yellow_detect, white_detect)
new_lanes = cv.cvtColor(lanes, cv.COLOR_HLS2BGR)
final = cv.cvtColor(new_lanes, cv.COLOR_BGR2GRAY)
img_blur = cv.bilateralFilter(final, 9, 120, 100)
canny_image = cv.Canny(img_blur, 100, 200)
pts,prediction , lane_classfied , rads = calculate_curves(canny_image)
rads[0] = rads[0]/5
rads[1] = rads[1]/3
color_blend = np.zeros_like(img).astype(np.uint8)
cv.fillPoly(color_blend, pts, (0,255, 0))
circles_image = draw_circles(color_blend)
circles_image = cv.cvtColor(circles_image, cv.COLOR_GRAY2BGR)
new_warp = cv.warpPerspective(color_blend, Hinv, (crop_img.shape[1], crop_img.shape[0]))
result = cv.addWeighted(img, 1, new_warp, 0.5, 0)
cv.putText(result, prediction, (50, 50),cv.FONT_HERSHEY_SIMPLEX, 1.5,(0,255,0),2, cv.LINE_AA)
scale_percent = 20 # percent of original size
width = int(circles_image.shape[1] * scale_percent / 100)
height = int(circles_image.shape[0] * scale_percent / 100)
dim = (width, height)
resized_circles_image = cv.resize(circles_image, dim, interpolation = cv.INTER_AREA)
# cv.imshow("re-sized circle",resized_circles_image)
width = int(img.shape[1] * scale_percent / 100)
height = int(img.shape[0] * scale_percent / 100)
dim = (width, height)
resized_original_image = cv.resize(img, dim, interpolation = cv.INTER_AREA)
# cv.imshow("re-sized original",resized_original_image)
canvas = np.ones([880,1880,3]).astype(np.uint8)
canvas[:720,:1280,:] = result[:,:,:]
canvas[0:144,1280:1536] = resized_original_image[:,:,:]
resized_lanes = cv.resize(lanes, dim, interpolation = cv.INTER_AREA)
canvas[0:144,1537:1793] = resized_lanes[:,:,:]
canvas[144:744,1280:1580] = lane_classfied[:600,:300,:]
canvas[144:744,1580:1880] = circles_image[:600,:300,:]
cv.putText(canvas, "(1)", (1300, 50),cv.FONT_HERSHEY_SIMPLEX, 1.5,(0,255,0),2, cv.LINE_AA)
cv.putText(canvas, "(2)", (1570, 50),cv.FONT_HERSHEY_SIMPLEX, 1.5,(0,255,0),2, cv.LINE_AA)
cv.putText(canvas, "(3)", (1300, 200),cv.FONT_HERSHEY_SIMPLEX, 1.5,(0,255,0),2, cv.LINE_AA)
cv.putText(canvas, "(4)", (1570, 200),cv.FONT_HERSHEY_SIMPLEX, 1.5,(0,255,0),2, cv.LINE_AA)
text_1 = '(1) : Original Image (2) : Image processing'
text_2 = '(3) : Warped and classified lanes (4) : Sliding windows curve detection'
curvature = 'Left Curvature = '+str(rads[0]) + ' '+ 'Right Curvature = ' + str(rads[1])
cv.putText(canvas,text_1, (50, 750),cv.FONT_HERSHEY_SIMPLEX, 1, (0, 255, 0), 1)
cv.putText(canvas,text_2, (50, 800),cv.FONT_HERSHEY_SIMPLEX, 1, (0, 255, 0), 1)
cv.putText(canvas,curvature, (50, 850),cv.FONT_HERSHEY_SIMPLEX, 1, (0, 255, 0), 1)
cv.imshow('canvas',canvas)
return canvas
def get_curve(img, leftx, rightx):
#create points from 0 to 399 to plot the curve
ploty = np.linspace(0, img.shape[0]-1, img.shape[0])
#max of y i.e 399
y_eval = np.max(ploty)
# Fit new polynomials to x,y in world space
left_fit_cr = np.polyfit(ploty, leftx, 2)
right_fit_cr = np.polyfit(ploty, rightx, 2)
# Calculate the new radii of curvature
left_curverad = ((1 + (2*left_fit_cr[0]*y_eval + left_fit_cr[1])**2)**1.5) / np.absolute(2*left_fit_cr[0])
right_curverad = ((1 + (2*right_fit_cr[0]*y_eval + right_fit_cr[1])**2)**1.5) / np.absolute(2*right_fit_cr[0])
#car position
car_pos = img.shape[1]/2
#calculating the lane center
l_fit_x_int = left_fit_cr[0]*img.shape[0]**2 + left_fit_cr[1]*img.shape[0] + left_fit_cr[2]
r_fit_x_int = right_fit_cr[0]*img.shape[0]**2 + right_fit_cr[1]*img.shape[0] + right_fit_cr[2]
lane_center_position = (r_fit_x_int + l_fit_x_int) /2
center = (car_pos - lane_center_position) / 10
return [left_curverad, right_curverad, center]
def draw_circles(frame):
img = frame.copy()
final = cv.cvtColor(img, cv.COLOR_BGR2GRAY)
# Filter noise
img_blur = cv.bilateralFilter(final, 9, 120, 100)
# Apply edge detection
img_edge = cv.Canny(img_blur, 100, 200)
corners = cv.goodFeaturesToTrack(img_edge,100,0.001,50)
corners = np.int0(corners)
# Iterate over the corners and draw a circle at that location
for i in corners:
x,y = i.ravel()
cv.circle(img_edge,(x,y),25,(255),1)
return img_edge
def calculate_curves(img_edge):
new_img = cv.warpPerspective(img_edge, H_matrix, (300, 600))
histogram = np.sum(new_img, axis=0)
out_img = np.dstack((new_img,new_img,new_img))*255
midpoint = np.int(histogram.shape[0]/2)
# Compute the left and right max pixels
leftx_ = np.argmax(histogram[:midpoint])
rightx_ = np.argmax(histogram[midpoint:]) + midpoint
left_lane_pos = leftx_
right_lane_pos = rightx_
image_center = int(new_img.shape[1]/2)
prediction = turn_predict(image_center, right_lane_pos, left_lane_pos)
window_height = np.int(new_img.shape[0]/num_windows)
nonzero = new_img.nonzero()
nonzeroy = np.array(nonzero[0])
nonzerox = np.array(nonzero[1])
leftx_p = leftx_
rightx_p = rightx_
left_lane_inds = []
right_lane_inds = []
# Step through the windows one by one
for window in range(num_windows):
# Identify window boundaries in x and y (and right and left)
win_y_down = new_img.shape[0] - (window+1)*window_height
win_y_up = new_img.shape[0] - window*window_height
win_x_left_down = leftx_p - margin
win_x_left_up = leftx_p + margin
win_x_right_down = rightx_p - margin
win_x_right_up = rightx_p + margin
window_pts = np.array([[win_x_left_up,win_y_down],
[win_x_right_up,win_y_down],
[win_x_left_down,win_y_up],
[win_x_right_down,win_y_up]])
# Identify the nonzero pixels in x and y within the window
good_left_inds = ((nonzeroy >= win_y_down) & (nonzeroy < win_y_up) & (nonzerox >= win_x_left_down) & (nonzerox < win_x_left_up)).nonzero()[0]
good_right_inds = ((nonzeroy >= win_y_down) & (nonzeroy < win_y_up) & (nonzerox >= win_x_right_down) & (nonzerox < win_x_right_up)).nonzero()[0]
# Append these indices to the list
left_lane_inds.append(good_left_inds)
right_lane_inds.append(good_right_inds)
# If found > minpix pixels, move to next window
if len(good_left_inds) > minpix:
leftx_p = np.int(np.mean(nonzerox[good_left_inds]))
if len(good_right_inds) > minpix:
rightx_p = np.int(np.mean(nonzerox[good_right_inds]))
# Concatenate the arrays of indices
left_lane_inds = np.concatenate(left_lane_inds)
right_lane_inds = np.concatenate(right_lane_inds)
# Extract left and right line pixel positions
leftx = nonzerox[left_lane_inds]
lefty = nonzeroy[left_lane_inds]
rightx = nonzerox[right_lane_inds]
righty = nonzeroy[right_lane_inds]
if leftx.size == 0 or rightx.size == 0 or lefty.size == 0 or righty.size == 0:
return
left_fit = np.polyfit(lefty, leftx, 2)
right_fit = np.polyfit(righty, rightx, 2)
ploty = np.linspace(0, new_img.shape[0]-1, new_img.shape[0] )
out_img[nonzeroy[left_lane_inds], nonzerox[left_lane_inds]] = [255, 0, 0]
out_img[nonzeroy[right_lane_inds], nonzerox[right_lane_inds]] = [0, 0, 255]
# Fit a second order polynomial to each
left_fitx = left_fit[0]*ploty**2 + left_fit[1]*ploty + left_fit[2]
right_fitx = right_fit[0]*ploty**2 + right_fit[1]*ploty + right_fit[2]
# Extracting points from fit
left_line_pts = np.array([np.transpose(np.vstack([left_fitx, ploty]))])
right_line_pts = np.array([np.flipud(np.transpose(np.vstack([right_fitx,ploty])))])
image_center = img_edge.shape[0]/2
pts = np.hstack((left_line_pts, right_line_pts))
pts = np.array(pts, dtype=np.int32)
rads = get_curve(new_img,left_fitx,right_fitx)
return pts , prediction , out_img , rads
def main():
cap = cv.VideoCapture('media/problem_3.mp4')
fourcc = cv.VideoWriter_fourcc('m','p','4', 'v')
out = cv.VideoWriter('output/problem_3_output.avi',cv.VideoWriter_fourcc('M','J','P','G'), 30, (1880,880))
while cap.isOpened():
success, frame = cap.read()
if success is False:
break
initial_image = frame
final = image_preprocessing(initial_image)
out.write(final)
cv.waitKey(1)
cap.release()
if __name__ == '__main__':
main()