def __init__(self, img):
self.image = img
self.HEIGHT = img.shape[0]
self.WIDTH = img.shape[1]
self.cte = 0
self.center_line = Line(self.WIDTH/2,self.HEIGHT,self.WIDTH/2,self.HEIGHT/2)
self.angle_steer = 0
self.speed = 10
self.dt = 0.1
self.yaw = 90 # phi(center,ox)
##initialize MPC
x_start = 140
y_start = 240
yaw = -1.4489
v = 10
#region of interest
self.vertices = np.array([[(int(0*img.shape[1]),img.shape[0]),(int(0*img.shape[1]),
int(0.15*img.shape[0])), (int(1.00*img.shape[1]), int(0.15*img.shape[0])), (
img.shape[1],img.shape[0])]], dtype=np.int32)
#img = cv2.cvtColor(self.image, cv2.COLOR_BGR2RGB)
self.final_img, self.left_line, self.right_line = LD.color_frame_pipeline([self.image], solid_lines=True)
self.speed_pub = rospy.Publisher("Team1_speed", Float32, queue_size = 10)
self.steer_pub = rospy.Publisher("Team1_steerAngle", Float32, queue_size = 10)
# start_time = time()
#self.sign_image, self.DIRECTION = TD.detect_sign(img)
# end_time = time()
# duration = end_time-start_time
# print(duration,"duration")
rate = rospy.Rate(30)
def open_zed():
print("Running...")
init = sl.InitParameters()
cam = sl.Camera()
if not cam.is_opened():
print("Opening ZED Camera...")
status = cam.open(init)
if status != sl.ERROR_CODE.SUCCESS:
print(repr(status))
exit()
runtime = sl.RuntimeParameters()
mat = sl.Mat()
key = ''
while key != 113: # for 'q' key
err = cam.grab(runtime)
if err == sl.ERROR_CODE.SUCCESS:
cam.retrieve_image(mat, sl.VIEW.VIEW_LEFT)
instance_image = cv2.imread(mat.get_data(), 0)
out_path = os.getcwd()
in_image = cv2.cvtColor(instance_image, cv2.COLOR_BGR2RGB)
plt.imshow("hi", in_image)
return
out_image = color_frame_pipeline([in_image], solid_lines=True)
cv2.imwrite(out_path, cv2.cvtColor(out_image, cv2.COLOR_RGB2BGR))
plt.imshow(out_image)
# print("hello")
plt.waitforbuttonpress()
plt.close('all')
# cv2.imshow("ZED", mat.get_data())
key = cv2.waitKey(5)
settings(key, cam, runtime, mat)
else:
key = cv2.waitKey(5)
cv2.destroyAllWindows()
cam.close()
print("\nFINISH")
figManager.resize(*figManager.window.maxsize())
# test on images
test_images_dir = join('data', 'test_images')
test_images = [
join(test_images_dir, name) for name in os.listdir(test_images_dir)
]
for test_img in test_images:
print('Processing image: {}'.format(test_img))
out_path = join('out', 'images', basename(test_img))
in_image = cv2.cvtColor(cv2.imread(test_img, cv2.IMREAD_COLOR),
cv2.COLOR_BGR2RGB)
out_image = color_frame_pipeline([in_image], solid_lines=True)
cv2.imwrite(out_path, cv2.cvtColor(out_image, cv2.COLOR_RGB2BGR))
if verbose:
plt.imshow(out_image)
plt.waitforbuttonpress()
plt.close('all')
# test on videos
test_videos_dir = join('data', 'test_videos')
test_videos = [
join(test_videos_dir, name) for name in os.listdir(test_videos_dir)
]
for test_video in test_videos:
print('Processing video: {}'.format(test_video))
import matplotlib.pyplot as plt
import cv2
import os
from os.path import join, basename
from lane_detection import color_frame_pipeline
RESIZE_H = 540
RESIZE_W = 960
WORKING_DIR = '/Users/dwang/self-driving-car/'
verbose = True
if __name__ == '__main__':
if verbose:
plt.ion()
test_images_dir = join(WORKING_DIR, 'project_1_lane_finding_basic/data/test_images/')
test_images = [join(test_images_dir, name) for name in os.listdir(test_images_dir)]
for test_img in test_images:
out_path = join(WORKING_DIR, 'out/images', basename(test_img))
print "processing image: {} to {}".format(test_img, out_path)
in_image = cv2.cvtColor(cv2.imread(test_img, cv2.IMREAD_COLOR), cv2.COLOR_BGR2RGB)
out_image = color_frame_pipeline([in_image], solid_lines=True)
cv2.imwrite(out_path, cv2.cvtColor(out_image, cv2.COLOR_RGB2BGR))
plt.close('all')
# test on images
test_images_dir = join('data', 'Our_test_images')
#test_images_dir = join('data', 'Our_test_images')
test_images = [
join(test_images_dir, name) for name in os.listdir(test_images_dir)
]
for test_img in test_images:
print('Processing image: {}'.format(test_img))
out_path = join('out', 'images', basename(test_img))
in_image = cv2.cvtColor(cv2.imread(test_img, cv2.IMREAD_COLOR),
cv2.COLOR_BGR2RGB)
out_image = color_frame_pipeline([in_image], solid_lines=True)
cv2.imwrite(out_path, cv2.cvtColor(out_image, cv2.COLOR_RGB2BGR))
if verbose:
plt.imshow(out_image)
plt.waitforbuttonpress()
plt.close('all')
# test on videos
test_videos_dir = join('data', 'test_videos')
test_videos = [
join(test_videos_dir, name) for name in os.listdir(test_videos_dir)
]
'''
for test_video in test_videos:
print('Processing video: {}'.format(test_video))
(h, w) = frame.shape[:2]
r = 320 / float(w)
dim = (320, int(h * r))
frame = cv2.resize(frame, dim, cv2.INTER_AREA)
# resize to 320 x 180 for wide frame
frame = frame[0:180, 0:320]
frame = cv2.rectangle(frame, (265, 105), (305, 55), (0, 255, 0), 2)
# crop for CNN model, i.e., traffic sign location
# can be adjusted based on camera angle
image = frame[55:105, 265:305]
frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
#frame = cv2.cvtColor(frame, cv2.COLOR_RGB2GRAY)
frame_buffer.append(frame)
blend_frame, lane_lines = color_frame_pipeline(frames=frame_buffer, \
solid_lines=True, \
temporal_smoothing=True)
# prepare the image to be classified by our deep learning network
image = cv2.resize(image, (28, 28), cv2.INTER_AREA)
#cv2.imshow("Image",image)
image = image.astype("float") / 255.0
image = img_to_array(image)
image = np.expand_dims(image, axis=0)
# classify the input image and initialize the label and
# probability of the prediction
# classify the input image
low_speed, RR, signal, high_speed, stop, yield1 = model.predict(image)[0]
label = "Not Stop"
