Key Frame Extraction-Based on Motion Information (Continued 3)

References motion-focusing Key Frame Extraction and Video Summarization for Lane surveillance system ICIP 2009

The train of thought and the previous articlesArticleThe specific ideas are as follows:

1). detected the moving target.

2) The detected moving targets are merged and overlapped to obtain a virtual key frame image.

The key lies in how to determine how many frames overlap to obtain a virtual key frame image.

Greedy is used in this article.Algorithm:

You can select one frame at a time to add frames in the overlapping area for fusion,

When the overlapping area of the selected frame and the current virtual image is greater than 95%,

The virtual key frame image is selected as the key frame.

Innovation: its ideas are novel.

Disadvantages:

This method can only be applied to scenarios with fixed velocity moving targets.