[−][src]Trait opencv::hub_prelude::CUDA_StereoSGM
The class implements the modified H. Hirschmuller algorithm HH08. Limitation and difference are as follows:
- By default, the algorithm uses only 4 directions which are horizontal and vertical path instead of 8. Set mode=StereoSGM::MODE_HH in createStereoSGM to run the full variant of the algorithm.
- Mutual Information cost function is not implemented. Instead, Center-Symmetric Census Transform with window size from Spangenberg2013 is used for robustness.
See also
cv::StereoSGBM
Required methods
pub fn as_raw_CUDA_StereoSGM(&self) -> *const c_void
[src]
pub fn as_raw_mut_CUDA_StereoSGM(&mut self) -> *mut c_void
[src]
Provided methods
pub fn compute(
&mut self,
left: &dyn ToInputArray,
right: &dyn ToInputArray,
disparity: &mut dyn ToOutputArray
) -> Result<()>
[src]
&mut self,
left: &dyn ToInputArray,
right: &dyn ToInputArray,
disparity: &mut dyn ToOutputArray
) -> Result<()>
Computes disparity map for the specified stereo pair
Parameters
- left: Left 8-bit or 16-bit unsigned single-channel image.
- right: Right image of the same size and the same type as the left one.
- disparity: Output disparity map. It has the same size as the input images. StereoSGM computes 16-bit fixed-point disparity map (where each disparity value has 4 fractional bits).
pub fn compute_with_stream(
&mut self,
left: &dyn ToInputArray,
right: &dyn ToInputArray,
disparity: &mut dyn ToOutputArray,
stream: &mut Stream
) -> Result<()>
[src]
&mut self,
left: &dyn ToInputArray,
right: &dyn ToInputArray,
disparity: &mut dyn ToOutputArray,
stream: &mut Stream
) -> Result<()>