Deep module and fitting module system and method for motion-based lane detection with multiple sensors

1. A method of lane detection for a non-transitory computer readable storage medium storing one or more programs, the one or more programs comprising instructions, which when executed by a computing device, cause the computing device to perform the following steps comprising:

generating a ground truth;

off-line training a lane detection algorithm by using the ground truth, the lane detection algorithm using parameters that express a lane marking in an arc;

on-line generating a predicted lane marking for a current view;

comparing the predicted lane marking against the ground truth; and

off-line refining the lane detection algorithm by using the lane template associated with the current view to generate an improved lane template used for improved lane detection of a next view;

wherein, a detected lane map is generated relative to a moving vehicle using the improved lane template and the predicted lane marking.

2. The method according to claim 1 , wherein generating a ground truth includes:

collecting data in an environment by using sensors that include an inertial measurement unit (IMU) module, a global positioning system (GPS) module and a mapping (MAP) module; and

generating a labeled lane marking by annotating a lane marking expressed in god's view.

3. The method according to claim 2 , wherein the IMU module provides information on at least one of a vehicle pose or a vehicle speed, and the global positioning system (GPS) module provides global position information.

4. The method according to claim 2 , wherein the MAP module is configured to create a map of its surroundings, and orient a vehicle itself within this map.

5. The method according to claim 1 , wherein on-line generating a predicted lane marking includes:

generating a hit-map image for the current view based on the lane detection algorithm; and

generating a fitted lane marking based on the hit-map image and a lane template that includes features of a view immediately previous to the current view.

6. The method according to claim 5 , wherein generating a fitted lane marking includes:

optimizing, based on priors or constraints, the lane template to obtain a local optimal.

7. The method according to claim 5 further comprising:

determining that a confidence level of the fitted lane marking is reasonable, using the parameters; and

outputting the fitted lane marking as a predicted lane marking.

8. The method according to claim 5 further comprising:

determining that a confidence level of the fitted lane marking is unreasonable, using the parameters; and

rejecting the fitted lane marking.

9. The method according to claim 1 , wherein off-line refining the lane detection algorithm includes:

adding additional ground truth data in the off-line training.

10. The method according to claim 1 further comprising:

on-line generating another predicted lane marking, using a refined lane detection algorithm.

11. A system for lane detection, the system comprising:

an internet server, comprising:

an I/O port, configured to transmit and receive electrical signals to and from a client device;

a memory;

one or more processing units; and

one or more programs stored in the memory and configured for execution by the one or more processing units, the one or more programs including instructions for:

generating a ground truth;

off-line training a lane detection algorithm by using the ground truth, the lane detection algorithm using parameters that express a lane marking in an arc;

on-line generating a predicted lane marking for a current view;

comparing the predicted lane marking against the ground truth; and

off-line refining the lane detection algorithm by using the lane template associated with the current view to generate an improved lane template used for improved lane detection of a next view;

wherein, a detected lane map is generated relative to a moving vehicle using the improved lane template and the predicted lane marking.

12. The system according to claim 11 , wherein generating a ground truth includes:

collecting data in an environment by using sensors that include an inertial measurement unit (IMU) module, a global positioning system (GPS) module and a mapping (MAP) module; and

generating a labeled lane marking by annotating a lane marking expressed in god's view.

13. The system according to claim 12 , wherein the IMU module provides information on at least one of a vehicle pose or a vehicle speed, and the global positioning system (GPS) module provides global position information.

14. The system according to claim 12 , wherein the MAP module is configured to create a map of its surroundings, and orient a vehicle itself within this map.

15. The system according to claim 11 , wherein on-line generating a predicted lane marking includes:

generating a hit-map image for the current view based on the lane detection algorithm; and

generating a fitted lane marking based on the hit-map image and a lane template that includes features of a view immediately previous to the current view.

16. The system according to claim 15 , wherein generating a fitted lane marking includes:

optimizing, based on priors or constraints, the lane template to obtain a local optimal.

17. The system according to claim 15 further comprising:

determining that a confidence level of the fitted lane is reasonable, using the parameters; and

outputting the fitted lane marking as a predicted lane marking.

18. The system according to claim 15 further comprising:

determining that a confidence level of the fitted lane is unreasonable, using the parameters; and

rejecting the fitted lane marking.

19. The system according to claim 11 , wherein off-line refining the lane detection algorithm includes:

adding additional ground truth data in the off-line training.

20. The system according to claim 11 further comprising:

on-line generating another predicted lane marking, using a refined lane detection algorithm.