Assessing visibility of a target object with autonomous vehicle fleet
A system uses a fleet of AVs to assess visibility of target objects. Each AV has a camera for capturing images of target objects. AVs provide the captured images, or visibility data derived from the captured images, to a remote system, which aggregates visibility data describing images captured across the fleet of AVs. The AVs also provide condition data describing conditions under which the images were captured, and the remote system aggregates the condition data. The remote system processes the aggregated visibility data and condition data to determine conditions under which a target object does not meet a visibility threshold.
1. A method for identifying a location for a target object, the method comprising:
capturing, by a camera mounted to an autonomous vehicle (AV), a plurality of images, each of the plurality of images having a field of view comprising a potential location for a target object, the plurality of images captured over a period of time;
determining whether the potential location for the target object is visible in one of the plurality of images;
in response to determining that the potential location for the target object is visible in one of the plurality of images, determining, based on the plurality of images, a duration of time during which the potential location is visible; and
in response to determining that the potential location is visible for at least a threshold period of time, transmitting the potential location for the target object to a remote system.
2. The method of claim 1 , further comprising:
identifying a potential location for a target object based on map data stored on the AV, the map data including locations and sizes of buildings in an environment of the AV.
3. The method of claim 2 , wherein the map data further includes data describing positions of windows and doors on at least a portion of the buildings in the environment of the AV, and the potential location comprises a portion of a building without a window or door.
4. The method of claim 2 , wherein the map data further includes a building type, and identifying the potential location for a target object comprises identifying a portion of a building of a particular building type.
5. The method of claim 1 , wherein determining whether the potential location for the target object is visible in one of the plurality of images comprises:
identifying, within the image, at least one boundary of the potential location for the target object;
determining an obstruction level of the potential location for the target object, the obstruction level describing an amount of the potential location that is obstructed from the camera by an obstructing object; and
determining whether the potential location for the target object is visible based on the obstruction level.
6. The method of claim 5 , further comprising
capturing, by a light detecting and ranging (lidar) sensor mounted to the AV, a point cloud describing at least the potential location for the target object and the obstructing object, and
classifying the obstructing object as an object type of a plurality of object types based on the captured point cloud.
7. The method of claim 1 , wherein determining a duration of time during which the potential location is visible comprises:
identifying a first image in which the potential location is visible, the first image captured at a first time;
identifying a second image in which the potential location is visible, the second image captured at a second time; and
determining the duration of time between the first time and the second time.
8. The method of claim 7 , further comprising determining that the potential location is visible in images of the plurality of images captured at times between the first time and the second time.
9. An autonomous vehicle (AV) to identify a location for a target object, the AV comprising:
a camera mounted to the AV, the camera to capture a plurality of images, each of the plurality of images having a field of view comprising a potential location for a target object, wherein the camera captures the plurality of images over a period of time;
a processor to:
receive the plurality of images captured by the camera;
determine whether the potential location for the target object is visible in one of the plurality of images; and
in response to determining that the potential location for the target object is visible in one of the plurality of images, determine, based on the plurality of images, a duration of time during which the potential location is visible; and
communications circuitry to transmit the potential location for the target object to a remote system in response to a determination that the potential location is visible for at least a threshold period of time.
10. The AV of claim 9 , the AV further comprising a memory to store map data, the map data including locations and sizes of buildings in an environment of the AV, and the processor further to identify a potential location for a target object based on the map data.
11. The AV of claim 10 , wherein the map data further includes data describing positions of windows and doors on at least a portion of the buildings in the environment of the AV, and the potential location comprises a portion of a building without a window or door.
12. The AV of claim 10 , wherein the map data further includes a building type, and the processor is further configured to identify the potential location for a target object by identifying a portion of a building of a particular building type.
13. The AV of claim 9 , wherein the processor, to determine whether the potential location for the target object is visible in one of the plurality of images, is configured to:
identify, within the image, at least one boundary of the potential location for the target object; and
determine an obstruction level of the potential location for the target object, the obstruction level describing an amount of the potential location that is obstructed from the camera by an obstructing object; and
determine whether the potential location for the target object is visible based on the obstruction level.
14. The AV of claim 13 , further comprising a light detecting and ranging (lidar) sensor mounted to the AV, the lidar sensor to capture a point cloud describing at least the potential location for the target object and the obstructing object; wherein the processor is further configured to classify the obstructing object as an object type of a plurality of object types based on the captured point cloud.
15. The AV of claim 9 , wherein the processor, to determine a duration of time during which the potential location is visible, is configured to:
identify a first image in which the potential location is visible, the first image captured at a first time;
identify a second image in which the potential location is visible, the second image captured at a second time; and
determine the duration of time between the first time and the second time.
16. A method for identifying a location for a target object, the method comprising:
receiving, from a plurality of AVs, data describing a potential location for a target object, the data from each of the plurality of AVs comprising whether the potential location for the target object was visible to a camera mounted to the AV, and, if the potential location for the target object was visible, a length of time during which the potential location for the target object was visible;
determining a proportion of the plurality of AVs for which the potential location for the target object was visible for at least a threshold period of time;
determining whether the proportion of the plurality of AVs exceeds a threshold proportion; and
in response, reporting the potential location for the target object as a suitable location for placing the target object.
17. The method of claim 16 , further comprising:
retrieving traffic flow data describing a level of traffic flow along a roadway passing by the potential location for the target object; and
reporting the potential location for the target object as a suitable location for placing the target object further in response to the level of traffic flow exceeding a threshold level.
18. The method of claim 16 , further comprising:
retrieving demographic data describing historical rideshare users traveling along a roadway passing by the potential location for the target object; and
reporting the demographic data with the potential location.
19. The method of claim 16 , further comprising providing, to the plurality if AVs, map data including locations and sizes of buildings in an environment of the AV, wherein an AV identifies the potential location for the target object based on the map data.
20. The method of claim 19 , wherein the map data further includes data describing positions of windows and doors on at least a portion of the buildings in the environment of the AV, and the potential location comprises a portion of a building without a window or door.