Question About RealSense D435i
-
Hi Bitxuyang RealSense 400 Series cameras such as D435i are Active Stereo. Structured Light is a different camera technology. The older retired SR300 RealSense camera model used a similar technology to Structured Light called Coded Light.
The Advanced Mode parameters of 400 Series cameras are controlled automatically with machine-learning algorithms due to the complex ways in which they interact with each other. For this reason the Advanced Mode settings are mostly undocumented, though users are free to perform experimentation with these settings to see how it affects the depth image.
Deep Neural Networks (DNN) can be used with RealSense 400 Series cameras, though they are not built into the camera and can instead be implemented with DNN programs such as the C++ / OpenCV example at the link below.
https://github.com/IntelRealSense/librealsense/tree/master/wrappers/opencv/dnn
The stereo algorithm of 400 Series cameras makes use of matching, though there are not a lot of details available as the algorithm is closed-source. The link below provides some information about the matching process.
https://github.com/IntelRealSense/librealsense/issues/3185#issuecomment-460540961
The Depth from Stereo guide provides more information about matching, including a Python script demonstrating SSD (Sum of Squared Differences) block-matching.
https://github.com/IntelRealSense/librealsense/blob/master/doc/depth-from-stereo.md
Please sign in to leave a comment.
Comments
1 comment