Dense mapping using range sensor multi-scanning and multi-view geometry from successive image frames

1. A system comprising:

a range sensor configured to receive signals reflected from an object in an environment and generate two or more successive scans of the environment at different times;

a camera configured to capture two or more successive camera images of the environment, wherein each of the two or more successive camera images of the environment is captured by the camera at a different location within the environment; and

processing circuitry configured to:

construct a spatial cost volume for the environment based on the two or more successive camera images;

determine a pixel-wise uncertainty for the spatial cost volume based on the two or more successive scans and the two or more successive camera images;

determine a rough an estimate of a depth of the object in the environment based on the spatial cost volume;

match points in the two or more successive scans to points in the two or more successive camera images;

determine a refined estimate of the depth of the object based on the rough estimate of the depth of the object and further based on matching the points on the successive range-sensor images to the successive camera images; and

generate a three-dimensional map of the environment based on the refined estimate of the depth of the object, the two or more successive scans, and the two or more successive camera images.

2. The system of claim 1 ,

wherein the object is a first object,

wherein the processing circuitry is further configured to estimate a depth of a second object in the environment based on the two or more successive scans and the two or more successive camera images, and

wherein the processing circuitry is configured to generate the three-dimensional map of the environment based on the refined estimate of the depth of the first object and the depth of the second object.

3. The system of claim 1 , wherein the processing circuitry is further configured to:

measure rotational movement of the camera relative to the range sensor; and

measure translational movement of the camera relative to the range sensor,

wherein the processing circuitry is configured to generate the three-dimensional map based on the two or more successive scans, the two or more successive camera images, the rotational movement of the camera, and the translational movement of the camera.

4. The system of claim 1 , wherein the range sensor comprises a radar sensor configured to transmit radar signals and receive the signals reflected from the objects in the environment.

5. The system of claim 1 , wherein the processing circuitry is further configured to perform simultaneous location and mapping based on the two or more successive camera images using the two or more successive scans as a depth constraint.

6. The system of claim 1 , wherein the processing circuitry is further configured to:

perform a first Fast Fourier Transform on the two or more successive scans to generate a third estimate of the depth of the object in the environment and a first estimate of the angle from the range sensor to the object;

perform a second Fast Fourier Transform on the two or more successive scans to generate a relative speed of the object; and

perform a third Fast Fourier Transform on the two or more successive scans to generate a second estimate of an angle from the range sensor to the object.

7. The system of claim 1 , wherein the processing circuitry is configured to determine the pixel-wise uncertainty for the spatial cost volume based on a rotational movement of the camera and a translational movement of the camera.

8. A method comprising:

receiving, by processing circuitry from a range sensor, two or more successive scans of an environment performed by the range sensor at different times, wherein the two or more successive scans represent information derived from signals reflected from an object in the environment;

receiving, by the processing circuitry, two or more successive camera images of the environment captured by the camera, wherein each of the two or more successive camera images of the object is captured by the camera at a different location within the environment;

constructing, by the processing circuitry, a spatial cost volume for the environment based on the two or more successive camera images;

determining, by the processing circuitry, a pixel-wise uncertainty for the spatial cost volume based on the two or more successive scans and the two or more successive camera images;

determining, by the processing circuitry, a rough estimate of a depth of the object in the environment based on the spatial cost volume;

matching points in the two or more successive scans to points in the two or more successive camera images;

determining a refined estimate of the depth of the object based on the rough estimate of the depth of the object and further based on matching the points on the successive range-sensor images to the successive camera images; and

generating, by the processing circuitry, a three-dimensional map of the environment based on the refined depth of the object, the two or more successive scans, and the two or more successive camera images.

9. The method of claim 8 , wherein the object is a first object, the method further comprising estimating a depth of a second object in the environment based on the two or more successive scans and the two or more successive camera images,

wherein generating the three-dimensional map of the environment is based on the refined estimate of the depth of the first object and the depth of the second object.

10. The method of claim 8 , further comprising:

measuring rotational movement of the camera relative to the range sensor; and

measuring translational movement of the camera relative to the range sensor,

wherein generating the three-dimensional map is based on the two or more successive scans, the two or more successive camera images, the rotational movement of the camera, and the translational movement of the camera.

11. The method of claim 8 , further comprising performing simultaneous location and mapping based on the two or more successive camera images using the two or more successive scans as a depth constraint.

12. The method of claim 8 , further comprising:

performing a first Fast Fourier Transform on the two or more successive scans to generate a third estimate of the depth of the object in the environment and a first estimate of the angle from the range sensor to the object,

performing a second Fast Fourier Transform on the two or more successive scans to generate a relative speed of the object; and

performing a third Fast Fourier Transform on the two or more successive scans to generate a second estimate of the angle from the range sensor to the object.

13. The method of claim 8 , wherein determining the pixel-wise uncertainty for the spatial cost volume is further based on a rotational movement of the camera and a translational movement of the camera.

14. A device comprising a computer-readable storage medium having executable instructions stored thereon, configured to be executable by processing circuitry of the device for causing the processing circuitry to:

receive, from a range sensor, two or more successive scans of an environment performed by the range sensor at different times, wherein the two or more successive scans represent information derived from signals reflected from an object in the environment;

receive, from a camera, two or more successive camera images of the environment, wherein each of the two or more successive camera images is captured by the camera at a different location within the environment;

construct a spatial cost volume for the environment based on the two or more successive camera images;

determine a pixel-wise uncertainty for the spatial cost volume based on the two or more successive scans and the two or more successive camera images;

determine a rough estimate of a depth of the object in the environment based on the spatial cost volume;

match points in the two or more successive scans to points in the two or more successive camera images;

determine a refined estimate of the depth of the object based on the rough estimate of the depth of the object and further based on matching the points on the successive range-sensor images to the successive camera images; and

generate a three-dimensional map of the environment based on the refined depth of the object, the two or more successive scans, and the two or more successive camera images.

15. The device of claim 14 , further comprising instructions for causing the processing circuitry to:

measure rotational movement of the camera relative to the range sensor; and

measure translational movement of the camera relative to the range sensor,

wherein the instructions to generate the three-dimensional map comprise instructions to generate the three-dimensional map based on the two or more successive scans, the two or more successive camera images, the rotational movement of the camera, and the translational movement of the camera.

16. The system of claim 1 , wherein the range sensor is configured to perform the two or more successive scans by transmitting signals across part or all of the environment for a first scan and then transmitting signals across part or all of the environment for a second scan.

17. The method of claim 8 , further comprising performing the two or more successive scans by causing the range sensor to transmit signals across part or all of the environment for a first scan and then to transmit signals across part or all of the environment for a second scan.

18. The device of claim 14 ,

wherein the object is a first object, and

wherein the device further comprises instructions for causing the processing circuitry to:

estimate a depth of a second object in the environment based on the two or more successive scans and the two or more successive camera images; and

generate the three-dimensional map of the environment based on the refined estimate of the depth of the first object and the depth of the second object.

19. The device of claim 14 , further comprising instructions for causing the processing circuitry to perform simultaneous location and mapping based on the two or more successive camera images using the two or more successive scans as a depth constraint.

20. The device of claim 14 , further comprising instructions for causing the processing circuitry to determine the pixel-wise uncertainty for the spatial cost volume based on a rotational movement of the camera and a translational movement of the camera.