Detection the cars rear tail-light

Discussion related to "under the hood" OpenMV topics.
skylin008
Posts: 45
Joined: Mon Mar 13, 2017 10:18 pm

Detection the cars rear tail-light

Postby skylin008 » Mon Aug 14, 2017 9:32 pm

We have a new project that needs to be detection in a variety of natural environments (including sunny, rainy, day, night, fog) need to detect the car's red taillights and its symmetry. When using the find_blob method, it is found that the influence of ambient light is very large, resulting in inaccurate detection. Is there a better algorithm to achieve this function, thanks!
User avatar
kwagyeman
Posts: 1175
Joined: Sun May 24, 2015 2:10 pm

Re: Detection the cars rear tail-light

Postby kwagyeman » Tue Aug 15, 2017 12:00 am

Find blobs is the best CV algorithm for this combined with get_histogram. I don't know how to do the math but you need to actively sample the image color space and look at the histogram and color statistics. Find blobs will just tell you an area in the image where the color is. But, you'll need to look at the color histogram to find what color to search for. You should look into how to parse LAB color histogram data to pull of red variances in an image. Note you can call get_histogram on ROIs in the image since you'll likely need to look at multiple regions of the image to get an idea for what's in it.
Nyamekye,

Return to “Technical Discussion”

Who is online

Users browsing this forum: No registered users and 1 guest