System and method for lateral vehicle detection
A system and method for lateral vehicle detection is disclosed. A particular embodiment can be configured to: receive lateral image data from at least one laterally-facing camera associated with an autonomous vehicle; warp the lateral image data based on a line parallel to a side of the autonomous vehicle; perform object extraction on the warped lateral image data to identify extracted objects in the warped lateral image data; and apply bounding boxes around the extracted objects.
1. A system comprising:
a data processor; and
a memory for storing a detection system, executable by the data processor, the detection system configured to:
receive lateral image data from a lateral-facing imaging device on a vehicle;
generate warped image data by warping the lateral image data based on linear edges or features of an object in the lateral image data such that the object in the warped image data becomes aligned with an orientation of the vehicle, wherein the warping includes aligning a bottom edge of the lateral image data with the linear edges or features of the object in the lateral image data, and wherein the bottom edge of the lateral image data is relative to the orientation of the vehicle; and
extract an extracted object from the object from the warped image data.
2. The system of claim 1 wherein the imaging device comprises a plurality of cameras, wherein the system is further configured to:
identify matching portions of extracted features from the warped image data from different ones of the plurality of cameras;
stitch together the warped image data based on the matching portions of the extracted features;
form a bounding box based on the matching portions of the extracted object; and
apply the bounding box around the extracted object in the warped image data.
3. The system of claim 2 further configured to:
apply portions of the bounding box on the warped image data from different ones of the plurality of cameras; and
stitch together the portions of the bounding box based on the matching portions of the extracted object.
4. The system of claim 1 wherein the imaging device is at least one laterally-facing camera associated with an autonomous vehicle, wherein the at least one laterally-facing camera is from a group consisting of: a forward laterally-facing camera and a backward laterally-facing camera.
5. The system of claim 1 , wherein the warped image data is further generated based on a line parallel to a side of the vehicle, and wherein the line parallel to the side of the vehicle is defined with configuration parameters corresponding to an installation orientation of the imaging device relative to the vehicle.
6. The system of claim 1 , wherein the warped image data is further generated based on a line parallel to a side of the vehicle, and wherein the line is parallel to linear edges or features of an object detected in the lateral image data.
7. The system of claim 1 , wherein the warped image data is further generated based on a line parallel to a side of the vehicle, and wherein the line parallel to the side of the vehicle is configured based on an installation of the imaging device and related configuration parameters, wherein the configuration parameters further correspond to an orientation of the bottom edge of the lateral image data.
8. A method comprising:
receiving lateral image data from a lateral-facing imaging device on a vehicle;
generating warped image data by warping the lateral image data based on linear edges or features of an object in the lateral image data such that the object in the warped image data becomes aligned with an orientation of the vehicle, wherein the warping includes aligning a bottom edge of the lateral image data with the linear edges or features of the object in the lateral image data, and wherein the bottom edge of the lateral image data is relative to the orientation of the vehicle; and
extracting an extracted object from the object from the warped image data.
9. The method of claim 8 wherein the warped image data creates a trapezoidal image.
10. The method of claim 8 further comprising:
applying a rectangular bounding box around the extracted object in the warped image data.
11. The method of claim 8 wherein the imaging device comprises a plurality of cameras, wherein the method further comprises:
identifying matching portions of extracted features from the warped image data from different ones of the plurality of cameras; and
stitching together the warped image data based on the matching portions of the extracted features, wherein the stitched image data forms a single contiguous image representing a stitched image from the plurality of cameras.
12. The method of claim 11 further comprising:
forming a bounding box based on the matching portions of the extracted object; and
applying the bounding box around the extracted object in the warped image data.
13. The method of claim 8 wherein the imaging device comprises a plurality of cameras, wherein the method further comprises:
identifying matching portions of extracted features from the warped image data from different ones of the plurality of cameras; and
combining the warped image data based on the matching portions of the extracted features, wherein the combined image data forms a single contiguous image representing a combined image from the plurality of cameras.
14. The method of claim 8 further comprising applying a bounding box around the extracted object in the warped image data.
15. A non-transitory machine-useable storage medium embodying instructions which, when executed by a machine, cause the machine to:
receive lateral image data from a lateral-facing imaging device on a vehicle;
generate warped image data by warping the lateral image data based on linear edges or features of an object in the lateral image data such that the object in the warped image data becomes aligned with an orientation of the vehicle, wherein the warping includes aligning a bottom edge of the lateral image data with the linear edges or features of the object in the lateral image data, and wherein the bottom edge of the lateral image data is relative to the orientation of the vehicle; and
extract an extracted object from the object from the warped image data.
16. The non-transitory machine-useable storage medium of claim 15 wherein the object in the warped image data is aligned with the orientation of the vehicle.
17. The non-transitory machine-useable storage medium of claim 15 wherein the imaging device comprises a plurality of cameras, wherein the non-transitory machine-useable storage medium embodying instruction further causes the machine to:
stitch together the warped image data from different ones of the plurality of cameras to form a single image; and
apply a bounding box around the extracted object in the single image.
18. The non-transitory machine-useable storage medium of claim 17 wherein the bounding box is trapezoidal.
19. The non-transitory machine-useable storage medium of claim 17 being further configured to modify the single image to be rectangular.
20. The non-transitory machine-useable storage medium of claim 15 being further configured to apply a two-dimensional (2D) bounding box around the extracted object in the warped image data.