Finding Lane Lines on the Road

Writeup Template

Reflection

1. Describe your pipeline. As part of the description, explain how you modified the draw_lines() function.

My pipeline consisted of 5 steps. First, I converted the images to grayscale, then I applied the Gaussian blur with a 3x3 filter on the output of the grayscale. After that I applied canny edge detection method to detect edges. Using Gaussian blur as a pre processing step to the canny method helps remove the unwanted edges(noisy) from the image. Then I used image masking and extract the region of interest from my image. Then I applied hough line transform on the resulted masked image to find out the line based on the votes we get for a particular grid in the hough space.

References:

1. https://dsp.stackexchange.com/questions/10057/gaussian-blur-standard-deviation-radius-and-kernel-size
2. https://alyssaq.github.io/2014/understanding-hough-transform/
3. https://medium.com/@esmat.anis/robust-extrapolation-of-lines-in-video-using-linear-hough-transform-edd39d642ddf
4. http://ottonello.gitlab.io/selfdriving/nanodegree/python/line%20detection/2016/12/18/extrapolating_lines.html
5. http://docs.opencv.org/3.0-beta/doc/py_tutorials/py_imgproc/py_houghlines/py_houghlines.html
6. http://www.pyimagesearch.com/2015/04/06/zero-parameter-automatic-canny-edge-detection-with-python-and-opencv/
7. https://github.com/vishalrangras/CarND-LaneLines-P1/blob/master/P1.ipynb

In order to draw a single line on the left and right lanes, I modified the draw_lines() function by taking the slope of the every line obtanined from the hough_lines() function. Slopes thus obtained can be distinguished based on whether they are positive or negative. Lines belong to the left lane have a positive slope, and the lines belong to the right lane have a negative slope. Hence based on the slope it is easy to differentiate left and right lane lines. Now to extrapolate the line points, I averaged the lane lines and obtained the top and botton line points for each left and right lanes.

2. Identify potential shortcomings with your current pipeline

One potential shortcoming would be what would happen when the pipeline runs on a video with curved lanes. For instance, the current pipelines is generating poor results on the video in the challeng section.

I think, another shortcoming could be to deal with occlusions(ex: vehicle in front) and illumination changes.

3. Suggest possible improvements to your pipeline

A possible improvement would be to use a logic that can detect the change in direction of the curve.

Another potential improvement could be to develop a logic that can detect transient conditions, such as changing lanes.