如何跟踪非骨架点或扩展Sketetal跟踪框架以添加加法关节 [英] How to Track Non Skeletal Point Or Extend the Sketetal Tracking Framework to Add Addition Joints

问题描述

例如，如果您想跟踪网球，以便您可以随时获得其位置（x，y，z）或扩展骨架跟踪框架。你将如何实现这一目标？

Say for example if you wanted to track a tennis ball such that you could get its position at all times (x,y,z) or extend the Skeletal Tracking Framework. How would you achieve this ?

推荐答案

身体追踪算法基于深度数据和人体的内置模型（我想我读了它的内置与mahcine学习技巧）。如果要跟踪其他对象，则必须构建自己的算法"来自scracth"。
即来自深度数据（颜色框也可以帮助）。

The body tracking algorithm is based on depth data and a built model of the human body (I think I read it's built with mahcine learning techniques). If you want to track a different object, you will have to build your own algorithm "from scracth" i.e. from the depth data (color frames can help too).

扩展Body跟踪并不容易，因为SDK算法没有打开，所以你将无法访问到"内部工作"当前可用的检测算法。

Extending the Body tracking is not easier, because the SDK algorithm is not open, so you will have no access to the "inner workings" of the currently available detection algorithm.

如果您想了解更多关于计算机视觉的信息，可以查看OpenCV文档中的教程。

If you want to know more about computer vision in general, you can have a look at the tutorials in OpenCV docummentation.

这篇关于如何跟踪非骨架点或扩展Sketetal跟踪框架以添加加法关节的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！