Fusing map reconstructions generated from multiple batches of data
Examples disclosed herein may involve a computing system that is operable to (i) receive image data captured from one or more devices, (ii) based on the image data, generate at least two batches of data corresponding to an area of a global map, wherein each batch of data comprises (a) a respective group of images from the received image data, and (b) one or more common images comprising one or more common visual features, (iii) generate a respective reconstruction of the area of the global map for each of the at least two batches of data, and (iv) fuse the respective reconstructions of the area of the global map using the one or more common visual features from the one or more common images.
1. A computer-implemented method comprising:
receiving image data captured from one or more devices;
based on the image data, generating at least two batches of data corresponding to an area of a global map, wherein each batch of data comprises:
a respective group of images from the received image data; and
one or more common images comprising one or more common visual features, wherein the one or more common images are common across the at least two batches of data;
generating a first reconstruction of the area of the global map based on a first batch of the at least two batches of data;
generating a second reconstruction of the area of the global map based on a second batch of the at least two batches of data; and
fusing the first and second reconstructions of the area of the global map using the one or more common visual features from the one or more common images.
2. The computer-implemented method of claim 1 , further comprising:
updating the global map with one or more of (i) one or more of the at least two batches of data when each of the first and second reconstructions of the area meets a predetermined quality threshold requirement, (ii) one or both of the first or second reconstructions of the area of the global map, or (iii) the fused first and second reconstructions of the area of the global map.
3. The computer-implemented method of claim 1 , wherein generating the first reconstruction of the area of the global map and generating the second reconstruction of the area of the global map comprises:
generating the first and second reconstructions independently in parallel.
4. The computer-implemented method of claim 1 , wherein:
generating the first reconstruction of the area of the global map based on the first batch of the at least two batches of data comprises:
optimizing the first reconstruction based on first visual features and first pose data generated from the image data; and
generating the second reconstruction of the area of the global map based on the second batch of the at least two batches of data comprises:
optimizing the second reconstruction based on second visual features and second pose data generated from the image data.
5. The computer-implemented method of claim 1 , wherein each of the at least two batches of data comprises one or more poses generated from the image data.
6. The computer-implemented method of claim 1 , wherein the image data comprises one or more image data collects.
7. The computer-implemented method of claim 1 , wherein generating the at least two batches of data comprises:
applying a set of constraints comprising one or more of (i) geographical radius, (ii) image size, (iii) number of images in each batch, or (iv) number of common images.
8. The computer-implemented method of claim 1 , wherein the one or more common visual features comprise one or more landmarks represented in the image data.
9. The computer-implemented method of claim 1 , wherein the image data is gathered at various times and in varying environmental conditions.
10. The computer-implemented method of claim 1 , wherein the image data is provided by one or more vehicles over one or more trajectories enabling each of the at least two batches of data to comprise a variety of image data corresponding to the same area of the global map.
11. The computer-implemented method of claim 1 , wherein each of the at least two batches of data comprises a visual representation of one or more trajectories of the one or more devices.
12. The computer-implemented method of claim 1 , further comprising:
performing bundle-adjustments on the first and second reconstructions to optimize pose data and localization landmarks between the at least two batches of data.
13. The computer-implemented method of claim 1 , wherein each of the first and second reconstructions of the area of the global map comprises a visualization of the area in a 3D space.
14. A computing system comprising:
at least one processor;
a non-transitory computer-readable medium; and
program instructions stored on the non-transitory computer-readable medium that are executable by the at least one processor such that the computing system is configured to perform functions comprising:
receiving image data captured from one or more devices;
based on the image data, generating at least two batches of data corresponding to an area of a global map, wherein each batch of data comprises:
a respective group of images from the received image data; and
one or more common images comprising one or more common visual features, wherein the one or more common images are common across the at least two batches of data;
generating a first reconstruction of the area of the global map based on a first batch of the at least two batches of data;
generating a second reconstruction of the area of the global map based on a second batch of the at least two batches of data; and
fusing the first and second reconstructions of the area of the global map using the one or more common visual features from the one or more common images.
15. The computing system of claim 14 , wherein the image data is gathered at various times and varying environmental conditions.
16. The computing system of claim 14 , wherein the image data is provided by one or more vehicles over one or more trajectories enabling each of the at least two batches of data to comprise a variety of image data corresponding to the same area of the global map.
17. The computing system of claim 14 , wherein:
generating the first reconstruction of the area of the global map based on the first batch of the at least two batches of data comprises:
optimizing the first reconstruction based on first visual features and first pose data generated from the image data; and
generating the second reconstruction of the area of the global map based on the second batch of the at least two batches of data comprises:
optimizing the second reconstruction based on second visual features and second pose data generated from the image data.
18. The computing system of claim 14 , wherein generating the first reconstruction of the area of the global map and generating the second reconstruction of the area of the global map comprises:
generating the first and second reconstructions independently in parallel.
19. A non-transitory computer-readable medium comprising program instructions stored thereon that, when executed by at least one processor of a computing system, cause the computing system to perform functions comprising:
receiving image data captured from one or more devices;
based on the image data, generating at least two batches of data corresponding to an area of a global map, wherein each batch of data comprises:
a respective group of images from the received image data; and
one or more common images comprising one or more common visual features, wherein the one or more common images are common across the at least two batches of data;
generating a first reconstruction of the area of the global map based on a first batch of the at least two batches of data;
generating a second reconstruction of the area of the global map based on a second batch of the at least two batches of data; and
fusing the first and second reconstructions of the area of the global map using the one or more common visual features from the one or more common images.
20. The non-transitory computer-readable medium of claim 19 , wherein generating the first reconstruction of the area of the global map and generating the second reconstruction of the area of the global map comprises
generating the first and second reconstructions independently in parallel.