This demo requires Bottlenose Stereo, it will not work with Bottlenose Mono.
This example assumes that:
- your camera is properly calibrated and the calibration parameters uploaded the camera.
- Use the Bottlenose Utility or checkout the calibration example to see how to upload your parameters.
- all the image quality related settings such as
exposure,gain, andCCMare set. Please useStereo VieweroreBusPlayerto configure the image quality to your liking.
The Python script shows how to:
- set keypoint parameters, only FAST is shown, but can be adapted for GFTT.
- set keypoint matching parameters
- enable chunk data transmission for sparse point cloud
The script will display the feature points that are detected in the left image that also
have valid 3d correspondences. Each valid frame will yield a ply and png files that can be viewed with a 3D
viewer such as MeshLab. For example,
python demo.py --mac <mac> --offsety1 <offset_from_calibration>
Set the following arguments to the demo.py file to change demo behavior.
| Parameter | Description |
|---|---|
--mac |
(optional) The MAC address of the camera. Assumes the first available camera if not specified. |
--max_keypoints |
(optional) Maximum number of keypoints to detect. Default 1000. |
--fast_threshold |
(optional) Keypoint threshold for the Fast9 algorithm. Default 4. |
--match_xoffset |
(optional) Matcher horizontal search range. Default 0 (use existing value) |
--match_yoffset |
(optional) Matcher vertical search range. Default 0 (use existing value) |
--offsety1 |
(optional) The Y offset of the right image. Default 440. |
--min_threshold |
(optional) Minimum threshold to accept a matched keypoint. Default 40. |
--ratio_threshold |
(optional) Ratio threshold. Default 1023 |
python demo.py <parameters>Back to Samples