ffmpeg去噪算法
2021/9/11 20:34:56
本文主要是介绍ffmpeg去噪算法,对大家解决编程问题具有一定的参考价值,需要的程序猿们随着小编来一起学习吧!
bm3d
ffmpeg -y -threads 16 -i input.mp4 -filter_complex "bm3d=sigma=5:block=4:bstep=2:group=1:estim=basic" output.mp4
ffmpeg -y -threads 16 -s 1920x1080 -i 5frame_1920x1080.yuv -filter_complex "bm3d=sigma=10:block=4:bstep=2:group=1:estim=basic" out.yuv
'sigma'
Set denoising strength. Default value is 1. Allowed range is from 0 to 999.9
http://underpop.online.fr/f/ffmpeg/help/bm3d.htm.gz
注意:ffmpeg需要4.3.1版本
nlm
ffmpeg -s $wx$h -pix_fmt yuv420p -i $input -vf nlmeans=h=2:range=3:temporal=1 $output
-vf nlmeans=h=2:range=3:temporal=1 为参数调节
官网
The filter is named nlmeans and has the following parameters:
h - averaging weight decay parameter (larger values give smoother videos).
range - spatial search range (default=3), should be odd number.
temporal - number of frames to include into search (default=2).
patchsize - pixel context region width (default=7, little need to change), should be odd number.
The default values (h=8, range=3, temporal=2) are a good starting point for the restoration of very noisy video (old VHS tapes).You may also try 10;5;3 for really noisy inputs or 6;3;1 for good quality inputs.
It is also interesting to note that one can simply apply the filter on interlaced material. The algorithm itself is robust enough to handle this nicely without artifacts.
https://www.dirk-farin.net/projects/nlmeans/index.html
这篇关于ffmpeg去噪算法的文章就介绍到这儿,希望我们推荐的文章对大家有所帮助,也希望大家多多支持为之网!
- 2024-05-15鸿蒙生态设备数量超8亿台
- 2024-05-13TiDB + ES:转转业财系统亿级数据存储优化实践
- 2024-05-09“2024鸿蒙零基础快速实战-仿抖音App开发(ArkTS版)”实战课程已上线
- 2024-05-09聊聊如何通过arthas-tunnel-server来远程管理所有需要arthas监控的应用
- 2024-05-09log4j2这么配就对了
- 2024-05-09nginx修改Content-Type
- 2024-05-09Redis多数据源,看这篇就够了
- 2024-05-09Google Chrome驱动程序 124.0.6367.62(正式版本)去哪下载?
- 2024-05-09有没有大佬知道这种数据应该怎么抓取呀?
- 2024-05-09这种运行结果里的10.100000001,怎么能最快改成10.1?