Tango 原始深度数据 - 更新? [英] Tango raw depth data - update?

问题描述

我从谷歌购买了支持探戈项目的 lenovo phab 2 pro.使用此设置，可以获得点云形式的深度数据.但这不是我需要的.我更喜欢以更原始的格式获取数据，例如 Kinect 可能获取的数据，其中图像平面的每个像素都分配了一个深度值.因此，我的问题是:phab2(或任何探戈设备)的深度数据是否可以以这样一种原始格式获得，其中每个像素都被分配了一个深度值?

I bought the lenovo phab 2 pro supporting the tango project from google. Using this setup, it it possible to obtain depth data in form of a pointcloud. But this is not what I need. I would prefer to obtain data in a more raw format like possible to be obtained by the Kinect, where each pixel of the imageplane is assigned a depth value. My question therefore: Is the depth data of the phab2 (or any tango device) possible to be obtained in such a raw format where each pixel is assigned a depth value?

我的研究使我发现了无数未解决的案例(在谷歌、stackoverflow 等中输入探戈原始数据或类似数据).这就是为什么我正在寻找有关此问题的任何更新，或者是否有人已经找到有关如何解决此问题的想法或有线索?另外，如果有人能告诉我如何，或者是否可以在不使用探戈的情况下访问深度相机的原始数据"，我会非常高兴?！目前我使用带有 tango 的 C-API 的 android sdk.

My research lead me to countless unsolved cases (typing tango raw data or similar in google, stackoverflow etc.). That is why I am looking for any updates on this issue or if anybody has already found an idea on how to solve this problem or has a lead? Also I would be greatful if somebody could tell me how, or if, it is possible to access the depth cameras "raw data" without using tango?! Currently I use the android sdk with tango's C-API.

推荐答案

没有直接通过 Tango API 可用的深度图像.实际上，Tango 考虑的原始数据是点云.为了获得深度图像，您必须将这些点投影到相机的图像平面上，并执行一定量的上采样以获得良好的结果.

There's no depth image directly available through Tango API. In fact, the raw data in Tango's consideration is point clouds. In order to get a depth image, you would have to project those points onto camera's image plane, and perform certain amount of upsampling to get a good result.

Tango 支持库 确实有几个 上采样方法.还有来自 Tango 的 rgb-depth-sync 示例代码展示了如何执行深度点投影.

Tango Support Library does have serveral upsample methods. There's also a rgb-depth-sync example code from Tango shows how to perform the depth points projection.

这篇关于Tango 原始深度数据 - 更新?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！