Method and apparatus for road hazard detection

An apparatus for road hazard detection, comprising:

• one or more sensors, the one or more sensor installed in a vehicle and configured to generate one or more sensor data, the one or more sensor data carrying traffic information that the vehicle encounters during driving;
• and
• processing circuitry configured to:
• acquire, via interface circuitry, the one or more sensor data from the one or more sensors;
• acquire, via the interface circuitry, traffic reference information from a database;
• compare the one or more sensor data with the traffic reference information;
• identify a road hazard based on a comparison result between a first sensor data from a first sensor and the traffic reference information;
• annotate other sensor data from other sensors to be associated with the road hazard, the other sensor data from other sensors being collected within a same time frame in which the first sensor data from the first sensor captures the road hazard;
• output, via the interface circuitry, the first sensor data from the first sensor that captures the road hazard and the other sensor data from the other sensors that are associated with the road hazard to a machine learning algorithm;
• train the machine learning algorithm based on the first sensor data from the first sensor that captures the road hazard and the other sensor data from the other sensors that are associated with the road hazard;
• and
• deploy the machine learning algorithm to identify a similar road hazard in a new event based on one or more sensor data that is acquired in the new event.

A non-transitory computer readable storage medium having instructions stored thereon that when executed by processing circuitry causes the processing circuitry to perform operations, the operations comprising:

• acquiring one or more sensor data generated by one or more sensors, the one or more sensor data carrying traffic information that the vehicle encounters during driving;
• acquiring traffic reference information from a database;
• comparing the one or more sensor data with the traffic reference information;
• identifying a road hazard based on a comparison result between a first sensor data from a first sensor and the traffic reference information;
• annotating other sensor data from other sensors to be associated with the road hazard, the other sensor data from other sensors being collected within a same time frame in which the first sensor data from the first sensor captures the road hazard;
• outputting the first sensor data from the first sensor that captures the road hazard and the other sensor data from other sensors that are associated with the road hazard to a machine learning algorithm;
• training the machine learning algorithm based on the first sensor data that captures the road hazard and the other sensor data that are associated with the road hazard;
• and
• deploying the machine learning algorithm to identify a similar road hazard in a new event based on one or more sensor data that is acquired in the new event.