DPV-SLAM

The DPVSLAM class implements DPV-SLAM using loop closure for globally consistent trajectory estimation and mapping.

class DPVSLAM()

use_local · boolean · default: true

If True, inference call is run on the local VM, else offloaded onto GRID-Cortex. Defaults to True.

calib · np.ndarray · default: "True"

The camera calibration matrix of shape $(4,)$. Defaults to np.array([320, 320, 320, 240]).

This model is currently not available via Cortex.

def run()

rgbimage · np.ndarray · required

The input RGB image of shape $(M,N,3)$.

Returns · np.ndarray

The predicted pose as a 1x6 array.

Inference Call

from grid.model.perception.slam.dpv_slam import DPVSLAM
car = AirGenCar()

# We will be capturing an image from the AirGen simulator 
# and run model inference on it.

img =  car.getImage("front_center", "rgb").data

model = DPVSLAM(use_local = True)
result = model.run(image=img)

License

This code is licensed under the MIT License.

Source

https://github.com/princeton-vl/DPVO