SDK 'software_device' which matching algorithm? #11926
-
|
Hello, when I use the 'software device' sample program, I can simulate a camera or transfer stereo-images to this program. Out of this stereo-images and the extrinsic parameters this program will give me a disparity map and a depth map - or am I wrong? best regards, |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment
-
|
Hi @HanzDieter You can transform depth representation to disparity representation by using the RealSense SDK instruction rs2::disparity_transform disparity_transform(true);, which is not used in the #10410 has C++ scripting for performing a depth to disparity transform with software-device. RealSense stereo depth algorithms are closed-source and so there is not much information available about them, though the Intel guide at the link below indicates that they are based on the SSD (Sum of Squared Differences) block-matching algorithm. https://github.com/IntelRealSense/librealsense/blob/master/doc/depth-from-stereo.md |
Beta Was this translation helpful? Give feedback.
Hi @HanzDieter You can transform depth representation to disparity representation by using the RealSense SDK instruction rs2::disparity_transform disparity_transform(true);, which is not used in the
rs-software-deviceexample program.#10410 has C++ scripting for performing a depth to disparity transform with software-device.
RealSense stereo depth algorithms are closed-source and so there is not much information available about them, though the Intel guide at the link below indicates that they are based on the SSD (Sum of Squared Differences) block-matching algorithm.
https://github.com/IntelRealSense/librealsense/blob/master/doc/depth-from-stereo.md