Search code examples
pythonopencvvideo-processingtensoryolo

background detection in frames


Is there any way to check for background before detecting the object by using yolo algorithm.For example if the background has white crossing lines,then draw the contour for person detection.Is this is possible?


Solution

  • You can solve your problem by detecting humans and zebra crossing, and filter out the boxes that are not close or doesn't intersect with the zebra crossing bounding box using some rules.

    Human detection can be done using Yolo. For detecting crosswalks, there is a lot of implementation on the internet, one of them is in this github repo