Rear obstacle detection and distance estimation

1. A method for rear obstacle detection using structure from motion, the method comprising:

identifying image features in a first frame corresponding to features in a second frame, wherein the first frame and the second frame comprise adjacent image frames captured by a rear-facing camera of a vehicle;

determining parameters for a non-planar motion model based on the image features;

generating a three-dimensional reconstruction of feature points of the image features based on the non-planar motion model;

determining a height of one or more objects based on corresponding feature points of the feature points;

determining a distance to the one or more objects having a height above a threshold;

determining camera motion based on the parameters for the non-planar motion model; and

notifying a driver or automated driving system of a presence of the one or more objects having the height above the threshold.

2. The method of claim 1 , further comprising identifying one or more feature points as part of a same object based on one or more of a pixel intensity or two-dimensional location within the first frame or second frame.

3. The method of claim 1 , further comprising determining a scale for location of the feature points in three-dimensional space based on feature points within a predefined ground area in the first frame and the second frame.

4. The method of claim 1 , further comprising classifying features as inliers and outliers, wherein determining the parameters comprises determining based on the inliers.

5. The method of claim 1 , further comprising performing local bundle adjustment on image features for improved accuracy.

6. A system for rear obstacle detection using structure from motion, the system comprising:

a monocular camera of a vehicle; and

a vehicle controller in communication with the rear-facing camera, wherein the vehicle controller is configured to:

obtain a series of image frames captured by the monocular camera during movement of the vehicle;

identify image features in a first frame corresponding to features in a second frame, wherein the first frame and the second frame comprise adjacent image frames in the series of image frames;

determine parameters for a non-planar motion model based on the image features;

generate a three-dimensional reconstruction of feature points of the image features based on the non-planar motion model;

determine a height of one or more objects based on corresponding feature points of the feature points;

determine a distance to the one or more objects having a height above a threshold;

determine camera motion based on the parameters for the non-planar motion model; and

notify a driver or automated driving system of a presence of the one or more objects having the height above the threshold.

7. The system of claim 6 , wherein the vehicle controller is further configured to identify one or more feature points as part of a same object based on one or more of a pixel intensity or two-dimensional location within the first frame or second frame.

8. The system of claim 6 , wherein the vehicle controller is further configured to determine a scale for location of the feature points in three-dimensional space based on feature points within a predefined ground area in the first frame and the second frame.

9. The system of claim 6 , wherein the vehicle controller is further configured to classify features as inliers or outliers, wherein determining the parameters comprises determining based on the inliers.

10. The method of claim 6 , wherein the vehicle controller is further configured to perform local bundle adjustment on image features for improved accuracy.

11. Non-transitory computer readable storage media storing instructions that, when executed by one or more processors, cause the one or more processors to:

identify image features in a first frame corresponding to features in a second frame, wherein the first frame and the second frame comprise adjacent image frames captured by a rear-facing camera of a vehicle;

determine parameters for a non-planar motion model based on the image features;

generate a three-dimensional reconstruction of feature points of the image features based on the non-planar motion model;

determine a height of one or more objects based on corresponding feature points of the feature points;

determine a distance to the one or more objects having a height above a threshold;

determine camera motion based on the parameters for the non-planar motion model; and

notify a driver or automated driving system of a presence of the one or more objects having the height above the threshold.

12. The non-transitory computer readable storage media of claim 11 , wherein the instructions further cause the one or more processors to identify one or more feature points as part of a same object based on one or more of a pixel intensity or two-dimensional location within the first frame or second frame.

13. The non-transitory computer readable storage media of claim 11 , wherein the instructions further cause the one or more processors to determine a scale for location of the feature points in three-dimensional space based on feature points within a predefined ground area in the first frame and the second frame.

14. The non-transitory computer readable storage media of claim 11 , wherein the instructions further cause the one or more processors to classify features as inliers and outliers, wherein determining the parameters comprises determining based on the inliers.