Systems and methods for updating a current base model comprising a plurality of images
Systems and methods for updating a current base model are provided. The systems include a controller programmed to obtain a current base model comprising a plurality of images, a 3D pointset, and geometric features, retrieve captured video data comprising a plurality of video frames with location data of a vehicle as the vehicle drives a route, localize the route of the vehicle and determine a pose of the vehicle using the video data from the vehicle and the current base model, identify one or more features, one or more objects, or both, present in the video data and not in the current base model, and perform a triangulation process, using the determined localization including location data of the vehicle and pose data of the vehicle, to position the identified features, the identified objects, or both, within the current base model to update the current base model.
1. A system comprising:
a controller programmed to:
obtain a current base model comprising a plurality of images;
retrieve captured video data comprising a plurality of video frames with location data of a vehicle as the vehicle drives a route;
localize the route of the vehicle and determine a pose of the vehicle, the pose of the vehicle comprising a location of the vehicle and an orientation of the vehicle at a same time, using the video data from the vehicle and the current base model;
identify one or more features, one or more objects, or both, present in the video data and not in the current base model; and
perform a triangulation process, using the determined localization including pose data of the vehicle, the pose data of the vehicle pertaining to the location of the vehicle and the orientation of the vehicle at the same time, to position the identified features, the identified objects, or both, within the current base model to update the current base model.
2. The system of claim 1 , wherein the route of the vehicle is localized and the pose of the vehicle is determined by:
finding a closest image among the plurality of images from the current base model to each of the plurality of video frames;
comparing the closest image to each of the plurality of video frames; and
performing the triangulation process based on a comparison of the closest images and each of the plurality of video frames.
3. The system of claim 2 , wherein the controller is further configured to:
identify feature points in the plurality of video frames;
identify feature points in the plurality of images in the current base model; and
compare the feature points in the plurality of video frames and the feature points in the plurality of images in the current base model to find the closest image to each of the plurality of video frames.
4. The system of claim 3 , wherein the controller is further configured to:
determine point correspondences between the closest image and each of the plurality of video frames to localize a location of the plurality of video frames in the current base model.
5. The system of claim 3 , wherein the feature points in the plurality of video frames, the feature points in the plurality of images, or both, comprise boundaries of geometry, corners of geometry, boundaries of signs, corners of signs, boundaries of roadway, corners of roadway, boundaries of obstacles, corners of obstacles, or combinations thereof.
6. The system of claim 1 , wherein the controller is further configured to:
obtain a plurality of spherical images to produce the current base model.
7. The system of claim 6 , wherein the controller is further configured to:
extract front images and back images from the plurality of spherical images.
8. The system of claim 7 , wherein the controller is further configured to:
identify feature points in two or more front images, two or more back images, or both;
determine point correspondences between the two or more front images, the two or more back images, or both;
triangulate feature points with other front images, other back images, or both; and
align the front images and the back images based on a result of triangulation of the feature points to produce the current base model.
9. The system of claim 8 , wherein each of the front images and the back images is a rectilinear image.
10. The system of claim 1 , wherein the plurality of images are taken by a first vehicle and the plurality of video frames are taken by a second vehicle different from the first vehicle.
11. The system of claim 1 , wherein the vehicle is a survey vehicle.
12. A method comprising:
obtaining a current base model comprising a plurality of images;
retrieving captured video data comprising a plurality of video frames with location data of a vehicle as the vehicle drives a route;
localizing the route of the vehicle and determine a pose of the vehicle, the pose of the vehicle comprising a location of the vehicle and an orientation of the vehicle at a same time, using the video data from the vehicle and the current base model;
identifying one or more features, one or more objects, or both, present in the video data and not in the current base model; and
performing a triangulation process, using the determined localization including pose data of the vehicle, the pose data of the vehicle pertaining to the location of the vehicle and the orientation of the vehicle at the same time, to position the identified features, the identified objects, or both, within the current base model to update the current base model.
13. The method of claim 12 , wherein the route of the vehicle is localized and the pose of the vehicle is determined by:
finding a closest image among the plurality of images from the current base model to each of the plurality of video frames;
comparing the closest image to each of the plurality of video frames; and
performing the triangulation process based on a comparison of the closest images and each of the plurality of video frames.
14. The method of claim 13 , further comprising:
identifying feature points in the plurality of video frames;
identifying feature points in the plurality of images in the current base model; and
comparing the feature points in the plurality of video frames and the feature points in the plurality of images in the current base model to find the closest image to each of the plurality of video frames.
15. The method of claim 14 , further comprising:
determine point correspondences between the closest image and each of the plurality of video frames to localize a location of the plurality of video frames in the current base model.
16. The method of claim 14 , wherein the feature points in the plurality of video frames, the feature points in the plurality of images, or both, comprise boundaries of geometry, corners of geometry, boundaries of signs, corners of signs, boundaries of roadway, corners of roadway, boundaries of obstacles, corners of obstacles, or combinations thereof.
17. The method of claim 12 , further comprising:
obtaining a plurality of spherical images to produce the current base model.
18. The method of claim 17 , further comprising:
extracting front images and back images from the plurality of spherical images.
19. The method of claim 18 , further comprising:
identifying feature points in two or more front images, two or more back images, or both;
finding point correspondences between the two or more front images, the two or more back images, or both;
triangulating feature points with other front images, other back images, or both; and
aligning the front images and the back images based on a result of triangulation of the feature points to produce the current base model.
20. The method of claim 19 , wherein each of the front images and the back images is a rectilinear image.