Detection and tracking of road-side pole-shaped static objects from LIDAR point cloud data

1. An advanced driver assistance system (ADAS) for a ground vehicle, the ADAS comprising:

a light detection and ranging (LIDAR) system configured to emit laser light pulses and capture reflected laser light pulses from the ground vehicle collectively forming three-dimensional (3D) LIDAR point cloud data for each of a plurality of frames; and

a controller configured to:

receive the 3D LIDAR point cloud data;

down-sample the 3D LIDAR point cloud data according to a user-configurable voxel size to reduce a size of the 3D LIDAR point cloud data and decrease a real-time computational load for at least one of increased processing throughput and increased speed of pole-shaped object detection and tracking;

filter the down-sampled 3D LIDAR point cloud data to remove data points corresponding to ground surface reflections;

divide the filtered 3D LIDAR point cloud data into a plurality of clusters that only include filtered 3D LIDAR point cloud data that is parallel to the ground vehicle's moving direction along a surface and within a configurable distance away from the ground vehicle;

project the plurality of clusters to a plane that is parallel to sides of the ground vehicle to generate an image;

detect a pole-shaped object in the image using an image-based object detection technique; and

track the pole-shaped object by deep sorting across multiple frames of the plurality of frames.

2. The ADAS of claim 1 , wherein the plurality of clusters are proximate to the sides of the vehicle.

3. The ADAS of claim 2 , wherein the controller is configured to divide the filtered 3D LIDAR point cloud data into the plurality of clusters according to at least one user-configurable cutting parameter.

4. The ADAS of claim 3 , wherein the at least one user-configurable cutting parameter includes at least one of thickness, stride, depth, and padding.

5. The ADAS of claim 1 , wherein the controller is configured to track the pole-shaped object by incorporating motion information based on a squared Mahalanobis distance between predicted Kalman states and newly arrived measurements.

6. The ADAS of claim 5 , wherein the controller is configured to track the pole-shaped object by inserting the pole-shaped object into a matching cascade that gives priority to more frequently seen objects to monitor a tracked position of the pole-shaped object.

7. The ADAS of claim 6 , wherein the controller is configured to track the pole-shaped object by predicting a position of the pole-shaped object using a standard Kalman filter with constant velocity motion and a linear observation model.

8. The ADAS of claim 7 , wherein the controller is configured to track the pole-shaped object based on a comparison between the tracked and predicted positions of the pole-shaped object.

9. The ADAS of claim 1 , wherein the controller does not utilize a radio detection and ranging (RADAR) system or a camera system in detecting and tracking the pole-shaped object.

10. The ADAS of claim 1 , wherein the pole-shaped object is one of a light or utility pole, a traffic sign, and a tree.

11. A method of detecting and tracking a pole-shaped object proximate to a ground vehicle, the method comprising:

receiving, by a controller of the vehicle and from a light detection and ranging (LIDAR) system of the vehicle, three-dimensional (3D) LIDAR point cloud data for each of a plurality of frames captured the LIDAR system after emitting laser light pulses and capturing reflected laser light pulses;

down-sampling, by the controller, the 3D LIDAR point cloud data according to a user-configurable voxel size to reduce a size of the 3D LIDAR point cloud data and decrease a real-time computational load for at least one of increased processing throughput and increased speed of pole-shaped object detection and tracking;

filtering, by the controller, the down-sampled 3D LIDAR point cloud data to remove data points corresponding to ground surface reflections;

dividing, by the controller, the filtered 3D LIDAR point cloud data into a plurality of clusters that only include filtered 3D LIDAR point cloud data that is parallel to the ground vehicle's moving direction along a surface and within a configurable distance away from the ground vehicle;

projecting, by the controller, the plurality of clusters to a plane that is parallel to sides of the ground vehicle to generate an image;

detecting, by the controller, a pole-shaped object in the image using an image-based object detection technique; and

tracking, by the controller, the pole-shaped object by deep sorting across multiple frames of the plurality of frames.

12. The method of claim 11 , wherein the plurality of clusters are proximate to the sides of the vehicle.

13. The method of claim 12 , wherein the dividing of the filtered 3D LIDAR point cloud data into the plurality of clusters is performed according to at least one user-configurable cutting parameter.

14. The method of claim 13 , wherein the at least one user-configurable cutting parameter includes at least one of thickness, stride, depth, and padding.

15. The method of claim 11 , wherein the tracking of the pole-shaped object comprises incorporating motion information based on a squared Mahalanobis distance between predicted Kalman states and newly arrived measurements.

16. The method of claim 15 , wherein the tracking of the pole-shaped object comprises inserting the pole-shaped object into a matching cascade that gives priority to more frequently seen objects to monitor a tracked position of the pole-shaped object.

17. The method of claim 16 , wherein the tracking of the pole-shaped object comprises predicting a position of the pole-shaped object using a standard Kalman filter with constant velocity motion and a linear observation model.

18. The method of claim 17 , wherein the tracking of the pole-shaped object is performed based on a comparison between the tracked and predicted positions of the pole-shaped object.

19. The method of claim 11 , wherein the controller does not utilize a radio detection and ranging (RADAR) system or a camera system in detecting and tracking the pole-shaped object.

20. The method of claim 11 , wherein the pole-shaped object is one of a light or utility pole, a traffic sign, and a tree.