Learning OpenCV 4 Computer Vision with Python 3 - Third Edition

Planning our MeanShift sample

For our first demonstration of MeanShift, we are not concerned with the approach to the initial detection of moving objects. We will use a naive approach that simply chooses the central part of the first video frame as our initial region of interest. (The user must ensure that an object of interest is initially located in the center of the video.) We will calculate a histogram of this initial region of interest. Then, in subsequent frames, we will use this histogram and the MeanShift algorithm to track the object.

Visually, the MeanShift demo will resemble many of the object detection samples that we have previously written. For every frame, we will draw a blue outline around the tracking rectangle, as shown ...