In this project, I used the skills that I learned in the second lecture to identify lane lines on the road.
I developed a pipeline on a series of individual images, and later applied the result to tewo video stream (really just a series of images).
Then I complete the optional challenge by implementing my version of daw_lines() function.
In [1]:
#importing some useful packages
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
import numpy as np
import cv2
import math
%matplotlib inline
In [2]:
# Global Variables
# Region-of-interest ofsets
top_x_offset = 0.45
top_y_offset = 0.60
bottom_x_offset = 0.07
# Stores the left and right lines from an image.
# Notice, need to clear before using it on new set of images (video).
right_lines = []
left_lines = []
In [3]:
import math
# Global Variables
# Region-of-interest ofsets
top_x_offset = 0.45
top_y_offset = 0.60
bottom_x_offset = 0.07
# Stores the left and right lines from an image.
# Notice, need to clear before using it on new set of images (video).
right_lines = []
left_lines = []
def grayscale(img):
"""Applies the Grayscale transform
This will return an image with only one color channel
but NOTE: to see the returned image as grayscale
(assuming your grayscaled image is called 'gray')
you should call plt.imshow(gray, cmap='gray')"""
return cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)
# Or use BGR2GRAY if you read an image with cv2.imread()
# return cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
def canny(img, low_threshold, high_threshold):
"""Applies the Canny transform"""
return cv2.Canny(img, low_threshold, high_threshold)
def gaussian_blur(img, kernel_size):
"""Applies a Gaussian Noise kernel"""
return cv2.GaussianBlur(img, (kernel_size, kernel_size), 0)
def region_of_interest(img, vertices):
"""
Applies an image mask.
Only keeps the region of the image defined by the polygon
formed from `vertices`. The rest of the image is set to black.
"""
#defining a blank mask to start with
mask = np.zeros_like(img)
#defining a 3 channel or 1 channel color to fill the mask with depending on the input image
if len(img.shape) > 2:
channel_count = img.shape[2] # i.e. 3 or 4 depending on your image
ignore_mask_color = (255,) * channel_count
else:
ignore_mask_color = 255
#filling pixels inside the polygon defined by "vertices" with the fill color
cv2.fillPoly(mask, vertices, ignore_mask_color)
#returning the image only where mask pixels are nonzero
masked_image = cv2.bitwise_and(img, mask)
return masked_image
def draw_lines(img, lines, color=[255, 0, 0], thickness=2):
for line in lines:
for x1,y1,x2,y2 in line:
cv2.line(img, (x1, y1), (x2, y2), color, thickness)
def hough_lines(img, rho, theta, threshold, min_line_len, max_line_gap):
"""
`img` should be the output of a Canny transform.
Returns an image with hough lines drawn.
"""
lines = cv2.HoughLinesP(img, rho, theta, threshold, np.array([]), minLineLength=min_line_len, maxLineGap=max_line_gap)
line_img = np.zeros((img.shape[0], img.shape[1], 3), dtype=np.uint8)
draw_lines(line_img, lines)
# draw_lines2(line_img, lines)
return line_img
# Python 3 has support for cool math symbols.
def weighted_img(img, initial_img, α=0.8, β=1., λ=0.):
"""
`img` is the output of the hough_lines(), An image with lines drawn on it.
Should be a blank image (all black) with lines drawn on it.
`initial_img` should be the image before any processing.
The result image is computed as follows:
initial_img * α + img * β + λ
NOTE: initial_img and img must be the same shape!
"""
return cv2.addWeighted(initial_img, α, img, β, λ)
def display_image(img, title):
plt.imshow(img, cmap='gray')
plt.suptitle(title)
plt.show()
def draw_lines2(img, lines, color=[255, 0, 0], thickness=2):
"""
This is my implementation of draw_lines() function
"""
threshold = 0.5
for line in lines:
for x1,y1,x2,y2 in line:
# Advoid divided by 0 exception
if x1 == x2:
continue
line_slope = ((y2-y1)/(x2-x1))
# only accept slopes >= threshold
if abs(line_slope) < threshold:
continue
if (line_slope >= threshold ): # Then its Right side
right_lines.append(line[0])
elif (line_slope < -threshold ): # Then its Left side
left_lines.append(line[0])
# Calculate Extrapolation based least-squares curve-fitting calculations of all the previous points
# Good link:
# https://ece.uwaterloo.ca/~dwharder/NumericalAnalysis/06LeastSquares/extrapolation/complete.html
right_lines_x = [x1 for x1, y1, x2, y2 in right_lines] + [x2 for x1, y1, x2, y2 in right_lines]
right_lines_y = [y1 for x1, y1, x2, y2 in right_lines] + [y2 for x1, y1, x2, y2 in right_lines]
# Calculate the slope (m) and the intercept (b), They are kept the same
# y = m*x + b
right_m = 1
right_b = 1
if right_lines_x:
right_m, right_b = np.polyfit(right_lines_x, right_lines_y, 1) # y = m*x + b
# collect left lines x and y sets for least-squares curve-fitting calculating
left_lines_x = [x1 for x1, y1, x2, y2 in left_lines] + [x2 for x1, y1, x2, y2 in left_lines]
left_lines_y = [y1 for x1, y1, x2, y2 in left_lines] + [y2 for x1, y1, x2, y2 in left_lines]
# Calculate the slope (m) and the intercept (b), They are kept the same
# y = m*x + b
left_m = 1
left_b = 1
if left_lines_x:
left_m, left_b = np.polyfit(left_lines_x, left_lines_y, 1)
# Calculate the y values
y_size = img.shape[0]
y1 = y_size
y2 = int(y_size*top_y_offset)
# Calculate the 4 points x values
right_x1 = int((y1-right_b)/right_m)
right_x2 = int((y2-right_b)/right_m)
left_x1 = int((y1-left_b)/left_m)
left_x2 = int((y2-left_b)/left_m)
# Graph the lines
if right_lines_x:
cv2.line(img, (right_x1, y1), (right_x2, y2), [255,0,0], 5)
if left_lines_x:
cv2.line(img, (left_x1, y1), (left_x2, y2), [255,0,0], 5)
Build your pipeline to work on the images in the directory "test_images"
You should make sure your pipeline works well on these images before you try the videos.
Build the pipeline and run your solution on all test_images. Make copies into the test_images_output directory, and you can use the images in your writeup report.
Try tuning the various parameters, especially the low and high Canny thresholds as well as the Hough lines parameters.
In [4]:
import os
os.listdir("test_images/")
Out[4]:
In [5]:
def process_image(image, display_images=False, export_images=False):
# Step 1. Convert the image to grayscale.
gray_image = grayscale(image)
# Step 2. Blur using Gaussian smoothing / blurring
# kernel_size = 5
blurred_image = gaussian_blur(gray_image, 5)
# Step 3. Use Canny Edge Detection to get a image of edges
# low_threshold = 50
# high_threshold = 150)
edged_image = canny(blurred_image, 50, 150)
# Step 4. Mask with a trapozoid the area of interest
y_size = image.shape[0]
x_size = image.shape[1]
tx = int(x_size * top_x_offset)
bx = int(x_size * bottom_x_offset)
ty = int(y_size * top_y_offset)
vertices = np.array( [[
(bx, y_size),# Bottom Left
(tx, ty), # Top Left
(x_size - tx, ty), # Top Right
(x_size - bx, y_size) # Bottom Right
]], dtype=np.int32 )
roi_img = region_of_interest(edged_image, vertices)
# Step 5. Run Hough Transformation on masked edge detected image
houghed_image = hough_lines(roi_img, 1, np.pi/180, 40, 30, 200)
# Step 6. Draw the lines on the original image
final_image = weighted_img(houghed_image, image)
if display_images:
display_image(image, "Original Image")
display_image(gray_image, "Grayscale Image")
display_image(blurred_image, "Gaussian Blured Image")
display_image(edged_image, "Canny Edge Detectioned Image")
display_image(roi_img, "Region of Interest Mapped Image")
display_image(houghed_image, "Hough Transformed Image")
display_image(final_image, "Final image")
if export_images:
mpimg.imsave("display_images_output" + "/" + "original_image.jpg", image, cmap='gray')
mpimg.imsave("display_images_output" + "/" + "gray_image.jpg", gray_image, cmap='gray')
mpimg.imsave("display_images_output" + "/" + "blurred_image.jpg", blurred_image, cmap='gray')
mpimg.imsave("display_images_output" + "/" + "edged_image.jpg", edged_image, cmap='gray')
mpimg.imsave("display_images_output" + "/" + "roi_img.jpg", roi_img, cmap='gray')
mpimg.imsave("display_images_output" + "/" + "houghed_image.jpg", houghed_image, cmap='gray')
mpimg.imsave("display_images_output" + "/" + "final_image.jpg", final_image, cmap='gray')
return final_image
In [6]:
final_image = process_image(mpimg.imread('test_images/solidWhiteRight.jpg'), display_images=True, export_images=True)
In [ ]:
in_directory = "test_images"
# Create a corresponding output directory
out_directory = "test_images_out"
if not os.path.exists(out_directory):
os.makedirs(out_directory)
# Get all images in input directory and store their names
imageNames = os.listdir(in_directory + "/")
for imageName in imageNames:
image = mpimg.imread(in_directory + "/" + imageName)
# Apply my Lane Finding Image Processing Algorithm on each image
resultImage = process_image(image)
# Save the result in the output directory
mpimg.imsave(out_directory + "/" + imageName, resultImage)
In [ ]:
# Import everything needed to edit/save/watch video clips
from moviepy.editor import VideoFileClip
from IPython.display import HTML
import imageio
imageio.plugins.ffmpeg.download()
In [ ]:
right_lines.clear()
left_lines.clear()
white_output = 'test_videos_output/solidWhiteRight.mp4'
clip1 = VideoFileClip("test_videos/solidWhiteRight.mp4")
white_clip = clip1.fl_image(process_image) #NOTE: this function expects color images!!
%time white_clip.write_videofile(white_output, audio=False)
In [ ]:
HTML("""
<video width="960" height="540" controls>
<source src="{0}">
</video>
""".format(white_output))
At this point, if you were successful with making the pipeline and tuning parameters, you probably have the Hough line segments drawn onto the road, but what about identifying the full extent of the lane and marking it clearly as in the example video (P1_example.mp4)? Think about defining a line to run the full length of the visible lane based on the line segments you identified with the Hough Transform. As mentioned previously, try to average and/or extrapolate the line segments you've detected to map out the full extent of the lane lines. You can see an example of the result you're going for in the video "P1_example.mp4".
Go back and modify your draw_lines function accordingly and try re-running your pipeline. The new output should draw a single, solid line over the left lane line and a single, solid line over the right lane line. The lines should start from the bottom of the image and extend out to the top of the region of interest.
Now for the one with the solid yellow lane on the left. This one's more tricky!
In [ ]:
right_lines.clear()
left_lines.clear()
yellow_output = 'test_videos_output/solidYellowLeft.mp4'
clip2 = VideoFileClip('test_videos/solidYellowLeft.mp4')
yellow_clip = clip2.fl_image(process_image)
%time yellow_clip.write_videofile(yellow_output, audio=False)
In [ ]:
HTML("""
<video width="960" height="540" controls>
<source src="{0}">
</video>
""".format(yellow_output))
In [ ]:
right_lines.clear()
left_lines.clear()
challenge_output = 'test_videos_output/challenge.mp4'
clip3 = VideoFileClip('test_videos/challenge.mp4')
challenge_clip = clip3.fl_image(process_image)
%time challenge_clip.write_videofile(challenge_output, audio=False)
In [ ]:
HTML("""
<video width="960" height="540" controls>
<source src="{0}">
</video>
""".format(challenge_output))