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?
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.
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 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.
来源：https://stackoverflow.com/questions/41572682/tango-raw-depth-data-update