基于树莓派的安全头盔视频稳像模块设计与实现

Design and Implementation of Video Stabilization Module for Safety Helmet Based on Raspberry Pi

安全头盔采集的视频具有与传统方式采集的视频不同的振动特性。在树莓派开发板上,基于经典的稀疏光流法,与基于特征点的方法相结合,设计与实现了一个安全头盔视频稳像模块,提高了输出视频图像的平稳度。重点讨论了如何结合树莓派开发板的特性,提高模块的运行速度的问题;从精简部分运算过程、波浪式并行执行、视频硬编码等角度,充分调用树莓派的运算能力,在实现视频稳像的同时减少了模块执行时间。

Compared with traditional methods,video captured by cameras mounted on the safety helmet has different vibration feature.A video stabilization module based on sparse optical flow algorithm combined with feature point method has been designed and implemented on the Raspberry Pi development board,and the stability of the output video is enhanced.The problem of how to accelerate the running speed of the module has been focused on according to the characteristics of Raspberry Pi.From the aspects including cancelling some of the calculation process,wave front style parallel operation and hardware based video coding,the calculation ability of Raspberry Pi is utilized.It efficiently helps to stabilize the output video while reduce the operation time of the module.