Skip to content

Overlay

This part of the pipeline is responsible for properly depicting the output. To do this, the original image is shown together with the detected lanes, threshold, and bird's eye view.

overlay_frames(undist, thresh, transform, img_fit)

Function which combines multiple images into one pretty output image

Source

Parameters:

Name Type Description Default
undist cv.Mat

Undistorted image blended with detected lane lines

 required
thresh cv.Mat

The thresholded binary image

 required
transform cv.Mat

The bird's eye view of the thresholded image

 required
img_fit cv.Mat

The bird's eye view with detected lane lines highlighted

 required

Returns:

Type Description
cv.Mat

The composed image containing
Source code in src/pipeline/overlay.py 
 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
def overlay_frames(undist: cv.Mat, thresh: cv.Mat, transform: cv.Mat, img_fit: cv.Mat):
    """Function which combines multiple images into one pretty output image

    !!! note "Source"

        - https://github.com/sidroopdaska/SelfDrivingCar/blob/master/AdvancedLaneLinesDetection/lane_tracker.ipynb

    Parameters
    ----------
    undist : cv.Mat
        Undistorted image blended with detected lane lines
    thresh : cv.Mat
        The thresholded binary image
    transform : cv.Mat
        The bird's eye view of the thresholded image
    img_fit : cv.Mat
        The bird's eye view with detected lane lines highlighted

    Returns
    -------
    cv.Mat
        The composed image containing
    """
    # Set the height and width, as well as offsets in pixels
    h, w = undist.shape[:2]
    thumb_ratio = 0.2
    thumb_h, thumb_w = int(thumb_ratio * h), int(thumb_ratio * w)
    off_x, off_y = 20, 15

    # Define a gray mask for separating the additional images from the original image
    mask = undist.copy()
    mask = cv.rectangle(mask, pt1=(0, 0), pt2=(w, thumb_h + 2 * off_y), color=(0, 0, 0), thickness=cv.FILLED)
    undist = cv.addWeighted(src1=mask, alpha=0.2, src2=undist, beta=0.8, gamma=0)

    # Add the top left thumbnail (thresholded binary image)
    thumb_binary = cv.resize(thresh, dsize=(thumb_w, thumb_h))
    thumb_binary = np.stack([thumb_binary] * 3, axis=2)
    undist[off_y : thumb_h + off_y, off_x : off_x + thumb_w, :] = thumb_binary

    # Add the center thumbnail (bird's eye view)
    thumb_birdeye = cv.resize(transform, dsize=(thumb_w, thumb_h))
    thumb_birdeye = np.stack([thumb_birdeye] * 3, axis=2)
    undist[off_y : thumb_h + off_y, int((w / 2) - (thumb_w / 2)) : int((w / 2) + (thumb_w / 2)), :] = thumb_birdeye

    # Add the top right thumbnail (bird's eye view with detected lane lines)
    thumb_img_fit = cv.resize(img_fit, dsize=(thumb_w, thumb_h))
    undist[off_y : thumb_h + off_y, w - off_x - thumb_w : w - off_x, :] = thumb_img_fit

    return undist