olpotkin / carnd-advanced-lane-lines Goto Github PK

The goals / steps of this project are the following:

• Compute the camera calibration matrix and distortion coefficients given a set of chessboard images.
• Apply a distortion correction to raw images.
• Use color transforms, gradients, etc., to create a thresholded binary image.
• Apply a perspective transform to rectify binary image ("birds-eye view").
• Detect lane pixels and fit to find the lane boundary.
• Determine the curvature of the lane and vehicle position with respect to center.
• Warp the detected lane boundaries back onto the original image.
• Output visual display of the lane boundaries and numerical estimation of lane curvature and vehicle position.

You're reading it!

The code for this step is contained in the code cells #2-3 of the IPython notebook located in "./CarND-Advanced-Lane-Lines.ipynb".

Step 1: Prepare "object points", which will be the (x, y, z) coordinates of the chessboard corners in the world.

Assuming the chessboard is fixed on the (x, y) plane at z=0, such that the object points are the same for each calibration image. Thus, objp is just a replicated array of coordinates, and objpoints will be appended with a copy of it every time algorithm successfully detects all chessboard corners in a test image.

imgpoints will be appended with the (x, y) pixel position of each of the corners in the image plane with each successful chessboard detection.

Step 2: Use the output objpoints and imgpoints to compute the camera calibration and distortion coefficients using the cv2.calibrateCamera() function.

Step 3: Apply this distortion correction to the test image using the cv2.undistort() function and obtained this result:

Pipeline contains combination of color and gradient thresholds to generate a binary image (thresholding steps at cell #5 through #7.

Step 1: Apply Sobel on x-axis (calculate directional gradient and apply gradient threshold)

Step 2: Apply Sobel on y-axis

Step 3: Apply Sobel x and y, compute the magnitude of the gradient and apply a threshold

Step 4: Apply Sobel x and y, computes the direction of the gradient and apply a threshold

Step 5: Combine the thresholds from steps 1-4

Step 6: Threshold color channel (S channel from HLS color space)

Step 7: Combine color and gradient thresholds (steps 5-6)

Results:

The code for perspective transform includes a function called warp(), which appears in code cell #12 of the IPython notebook. The warp() function takes as inputs an image (thresholded). The source and destination points were hardcoded in the following manner:

src = np.float32([[180, img.shape[0]], [575, 460], [705, 460], [1150, img.shape[0]]])

dst = np.float32([[320, img.shape[0]], [320, 0], [960, 0], [960, img.shape[0]]])

This resulted in the following source and destination points:

Draw the src and dst points onto a test image and its warped counterpart to verify that perspective transform works as expected. Lines appear parallel in the warped image:

Example with warped and masked image:

This step implemented in cell #13 in the function detect_lines()

Steps to fit lane lines with a 2nd order polynomial:

Step 1: Find the starting point for the left and right lines (take a histogram of the bottom half of the masked image)

Step 2: Set the width of the windows +/- margin (= 100), set minimum number of pixels found to recenter window (= 50)

Step 3: If lines are not detected - step through the windows one by one

Step 4: Extract left and right line pixel positions and Fit a second order polynomial to each (using np.polyfit

Step 5: Generate x and y values for plotting

This step implemented in cell #14 in the function rad_and_offset

Fit the left and right lines with a second order polynomial shown in green:

This step implemented in cell #15 in the function draw_res().

Here is an example of result on a test image:

Here's a link to my video result

Here's link to my video result on YouTube

Pipeline becames relatively nonstable in cases of very fast turns and difficult light conditions (experimental way to fix this issue is to use high light sensitive camera with high fps values).

• Drawing procedure could be improved by adding stabilization features

• More experiments with threshold values and color spaces (and different combination of both) could be performed

• Include additional information on each frame (statistics, etc.)