Intelligent systems and methods for mobilizing resources and assessing vehicle damage using aerial sensor data
A system and method for capturing and processing sensor data, such as aerial images of structures, to assess vehicular damage is disclosed. The system comprises a computing system used to obtain images of selected locations where large groups of vehicles were disposed during a natural or man-made disaster. The images are used to automatically determine a damage score for the vehicles. In addition, the imagery can be fed into a damage classifier that automatically classifies the degree of damage. The system and method may be used to quickly assess damage of vehicles in a disaster area and deploy resources to provide timely and swift assistance.
1 . A method of mobilizing resources for disaster-impacted areas using one or more aerial drones, the method comprising:
sending a drone dispatch signal to the one or more aerial drones, the drone dispatch signal causing the one or more aerial drones to deploy to a first region;
each of the one or more aerial drones including (1) an altitude and heading reference system that generates orientation and altitude data describing a position of the respective aerial drone, (2) a GPS receiver that generates GPS location data describing a position of the respective aerial drone, and (3) an image capture and processing system that generates image data;
receiving at a first time, at a damage assessment system, first sensor data for the first region, the first sensor data including image data of aerial views of a first group of vehicles;
the first sensor data being received from the one or more aerial drones, and the first sensor data further including the orientation and altitude data, and the GPS location data generated by the aerial drone;
passing the first sensor data through a machine learning damage classification model of the damage assessment system;
generating, with the machine learning damage classification model and based on the first sensor data that includes each of the image data, the orientation and altitude data, and the GPS location data, a damage report indicating a first percentage of the first group of vehicles that has been damaged; and
transmitting, from the damage assessment system and based on the damage report, a control signal to a damage response module requesting deployment of a first allocation of a first type of resource to the first region.
2 . The method of claim 1 , further comprising:
receiving, prior to the first time and at the damage assessment system, a first data indicating a disaster has occurred in the first region;
determining, at the damage assessment system, that the first data corresponds to a triggering event; and
transmitting, from the damage assessment system and to a drone repository, a request for the first sensor data to be captured.
3 . The method of claim 2 , wherein the method further comprises receiving layout data describing the first region, and the drone dispatch signal includes an aerial path of the one or more drones that ensures that the one or more drones will fly over an entirety of the first region based on the layout data.
4 . The method of claim 3 , wherein the machine learning damage classification model is trained on training image data that includes (1) images of empty parking lots, (2) images of parking lots with vehicles without damage, and (3) images of parking lots with vehicles having damage, in order to identify vehicles having damage in the first sensor data.
5 . The method of claim 1 , further comprising:
receiving, prior to the first time and at the damage assessment system, a first data indicating a disaster has occurred in the first region;
determining, at the damage assessment system, that the first data corresponds to a triggering event; and
transmitting, from the damage assessment system and to an aerial image database, a request for the first sensor data to be shared with the damage assessment system.
6 . The method of claim 1 , wherein the first region includes a parking lot where the first group of vehicles are parked.
7 . The method of claim 1 , wherein the first sensor data includes images providing top-down views of the first group of vehicles.
8 . The method of claim 1 , wherein the first group of vehicles includes a first vehicle, and the method further comprises:
receiving, prior to the first time and at the damage assessment system, a first insurance claim for damage to the first vehicle;
determining, at the damage assessment system, that the first claim corresponds to a triggering event; and
transmitting from the damage assessment system, to one of a drone repository and an aerial image database, a request for the first sensor data.
9 . The method of claim 8 , further comprising determining, at the damage assessment system the first vehicle is associated with a first damage level, wherein the first type of resource includes an automatic partial payout of the first insurance claim.
10 . The method of claim 1 , wherein the first type of resource includes one of an automated partial payout of insurance claims submitted on behalf of one or more vehicles in the group of vehicles and provisioning of a rental car for an insured member associated with a vehicle in the group of vehicles.
11 . A method of mobilizing resources for disaster-impacted areas using one or more aerial drones, the method comprising:
sending a drone dispatch signal to the one or more aerial drones, the drone dispatch signal causing the one or more aerial drones to deploy to a first region;
each of the one or more aerial drones including (1) an altitude and heading reference system that generates orientation and altitude data describing a position of the respective aerial drone, (2) a GPS receiver that generates GPS location data describing a position of the respective aerial drone, and (3) an image capture and processing system that generates image data;
receiving at a first time, at a damage assessment system, first sensor data for the first region, the first sensor data including image data of aerial views of a first group of vehicles, the first group of vehicles including a first vehicle;
the first sensor data being received from the one or more aerial drones, and the first sensor data further including the orientation and altitude data, and the GPS location data generated by the aerial drone;
receiving, prior to the first time and at the damage assessment system, a first insurance claim for damage to the first vehicle;
passing the first sensor data through a machine learning damage classification model of the damage assessment system;
generating, with the machine learning damage classification model and based on the first sensor data that includes each of the image data, the orientation and altitude data, and the GPS location data, a damage report indicating the first vehicle is associated with a first damage score; and
transmitting, from the damage assessment system and based on the damage report, a control signal to a damage response module that causes an automated first partial payout to be generated in response to the first insurance claim.
12 . The method of claim 11 , wherein the machine learning damage classification model is trained on training image data that includes (1) images of empty parking lots, (2) images of parking lots with vehicles without damage, and (3) images of parking lots with vehicles having damage, in order to identify vehicles having damage in the first sensor data.
13 . The method of claim 11 , wherein the first insurance claim includes information describing a current location of the first vehicle, and the information is used to identify the first vehicle in the first sensor data.
14 . A system for mobilizing resources for disaster-impacted areas, the system comprising a processor and machine-readable media including instructions which, when executed by the processor, cause the processor to:
send a drone dispatch signal to the one or more aerial drones, the drone dispatch signal causing the one or more aerial drones to deploy to a first region;
each of the one or more aerial drones including (1) an altitude and heading reference system that generates orientation and altitude data describing a position of the respective aerial drone, (2) a GPS receiver that generates GPS location data describing a position of the respective aerial drone, and (3) an image capture and processing system that generates image data;
receive at a first time, at a damage assessment system, a first sensor data for the first region, the first sensor data including image data of aerial views of a first group of vehicles;
the first sensor data being received from the one or more aerial drones, and the first sensor data further including the orientation and altitude data, and the GPS location data generated by the aerial drone;
pass the first sensor data through a machine learning damage classification model of the damage assessment system;
generate, with the machine learning damage classification model and based on the first sensor data that includes each of the image data, the orientation and altitude data, and the GPS location data, a damage report indicating a first percentage of the first group of vehicles that has been damaged; and
transmit, from the damage assessment system and based on the damage report, a control signal to a damage response module requesting deployment of a first allocation of a first type of resource to the first region.
15 . The system of claim 14 , wherein the instructions further cause the processor to:
receive, prior to the first time and at the damage assessment system, a first data indicating a disaster has occurred in the first region;
determine, at the damage assessment system, that the first data corresponds to a triggering event; and
transmit, from the damage assessment system and to a drone repository, a request for the first sensor data to be captured.
16 . The system of claim 14 , wherein the instructions further cause the processor to:
receive, prior to the first time and at the damage assessment system, a first data indicating a disaster has occurred in the first region;
determine, at the damage assessment system, that the first data corresponds to a triggering event; and
transmit, from the damage assessment system and to an aerial image database, a request for the first sensor data to be shared with the damage assessment system.
17 . The system of claim 14 , wherein the machine learning damage classification model is trained on training image data that includes (1) images of empty parking lots, (2) images of parking lots with vehicles without damage, and (3) images of parking lots with vehicles having damage, in order to identify vehicles having damage in the first sensor data.
18 . The system of claim 14 , wherein the first group of vehicles includes a first vehicle, and the instructions further cause the processor to:
receive, prior to the first time and at the damage assessment system, a first insurance claim for damage to the first vehicle;
determine, at the damage assessment system, that the first claim corresponds to a triggering event; and
transmit from the damage assessment system, to one of a drone repository and an aerial image database, a request for the first sensor data.
19 . The system of claim 18 , wherein the instructions further cause the processor to determine, at the damage assessment system the first vehicle is associated with a first damage level, wherein the first type of resource includes an automatic partial payout of the first insurance claim.
20 . The system of claim 14 , wherein the first type of resource includes one of a automated partial payout of insurance claims submitted on behalf of one or more vehicles in the group of vehicles and provisioning of a rental car for an insured member associated with a vehicle in the group of vehicles.