Closed wayy12 closed 8 years ago
Hi @wayy12 - no, librealsense is very minimal and doesn't contain any high-level APIs for computer vision. Not sure what implementation is in the RealSense SDK. Fortunately, none of them are super complicated to pull from sources like OpenCV or PCL. In general, your best bet is to start looking at bilateral, median, or guided (depth + rgb) filters. Usually, most of the complexity in them comes from their parallelized implementations (OpenCL, CUDA, etc) since they're not the speediest operations to do in real-time.
I noted that :in windows RSSDK provides a API
virtual PXCPhoto* PXCAPI EnhanceDepth(const PXCPhoto *photo, DepthFillQuality depthQuality)= 0
Then I thought whether the librealsense provide the similar API to enhance the depth image quality by filling holes and denoising?
I only found this method to get better depth data: dev.enable_stream(rs::stream::depth, rs::preset::best_quality); If I use this API to get depth data,Can I judge it as EnhanceDepth?
TKS!