# Generating action shot image using open cv

**URL:** <https://forum.opencv.org/t/generating-action-shot-image-using-open-cv/20689>\
**Category:** Uncategorized\
**Created:** [March 31, 2025, 12:54pm UTC](https://forum.opencv.org/t/generating-action-shot-image-using-open-cv/20689 "2025-03-31T12:54:24Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![rndzzz](https://avatars.discourse-cdn.com/v4/letter/r/97f17d/32.png) [@rndzzz](https://forum.opencv.org/u/rndzzz)\
**Post date:** [March 31, 2025, 12:54pm UTC](https://forum.opencv.org/t/generating-action-shot-image-using-open-cv/20689/1 "2025-03-31T12:54:24Z")

</div>

I am building a project that can generate an actionshot image using burst shots or extracted frames from a video. The problem is when I use image stitching (feature extraction, matching, homography, warping and blending), the layers of the images in the output is transparent or has “ghosting effects” on the subject. Are there other ways to generate like a panoramic output without using the stitcher class? Thanks

---

<div class="post-metadata">

**Author:** ![berak](https://avatars.discourse-cdn.com/v4/letter/b/85f322/32.png) [@berak](https://forum.opencv.org/u/berak)\
**Post date:** [March 31, 2025, 2:28pm UTC](https://forum.opencv.org/t/generating-action-shot-image-using-open-cv/20689/2 "2025-03-31T14:28:11Z")

</div>

> [@rndzzz](#):
>
> the layers of the images in the output is transparent or has “ghosting effects” on the subject

imposible to see from your description. please add example code / image(s), ty.

---

<div class="post-metadata">

**Author:** ![berak](https://avatars.discourse-cdn.com/v4/letter/b/85f322/32.png) [@berak](https://forum.opencv.org/u/berak)\
**Post date:** [March 31, 2025, 3:26pm UTC](https://forum.opencv.org/t/generating-action-shot-image-using-open-cv/20689/4 "2025-03-31T15:26:12Z")

</div>

code, input ? we still have no idea, what you’re doing

---

<div class="post-metadata">

**Author:** ![crackwitz](https://sea2.discourse-cdn.com/flex020/user_avatar/forum.opencv.org/crackwitz/32/14_2.png) [@crackwitz](https://forum.opencv.org/u/crackwitz)\
**Post date:** [March 31, 2025, 5:12pm UTC](https://forum.opencv.org/t/generating-action-shot-image-using-open-cv/20689/5 "2025-03-31T17:12:52Z")

</div>

based on that deleted picture, you seem to want a composite of the _changes_.

you’ll need a lot of control over the compositing step. you’ll need to determine what’s background and what’s foreground/action from taking the distribution of “fragments” (data from each image) for each pixel, and hopefully there’s just one image that is a clear outlier in the distribution. that’s the one with the “action”. you take that pixel.

you’ll need to fiddle with thresholds. picking the outlier indiscriminately will increase noise in entirely background pixels.