Tracking vehicles in a warehouse environment

A system, comprising:

• a vehicle;
• a first stereoscopic camera affixed to the vehicle;
• a second stereoscopic camera affixed to the vehicle;
• and
• a computing device communicatively coupled to each of the first and second stereoscopic cameras, the computing device configured to perform operations comprising:
• recognizing a first object that is represented in first stereoscopic image data based on one or more stereoscopic images received from a first stereoscopic camera that has been affixed to a vehicle;
• recognizing a second object that is represented in second stereoscopic image data based on one or more stereoscopic images received from a second stereoscopic camera that has been affixed to the vehicle;
• identifying respective representations of the first object and the second object in a spatial model that tracks, for each of a plurality of objects, a corresponding location in an environment and a level of confidence that the corresponding location is the object's actual location;
• determining a first location of the first object in the environment and a first level of confidence that the first location is the first object's actual location, according to the spatial model;
• determining a second location of the second object in the environment and a second level of confidence that the second location is the second object's actual location, according to the spatial model;
• and
• in response to determining that the first level of confidence for the first object's location is greater than the second level of confidence for the second object's location:
• determining a relative position between the vehicle and the first object, based on a portion of the first stereoscopic image data that represents the first object;
• and
• determining a location of the vehicle in the environment, based on the determined first location of the first object in the environment, and the determined relative position between the vehicle and the first object.

A non-transitory computer-readable storage medium having instructions stored thereon which, when executed by one or more processors, cause the one or more processors to perform operations comprising:

• recognizing a first object that is represented in first stereoscopic image data based on one or more stereoscopic images received from a first stereoscopic camera that has been affixed to a vehicle;
• recognizing a second object that is represented in second stereoscopic image data based on one or more stereoscopic images received from a second stereoscopic camera that has been affixed to the vehicle;
• identifying respective representations of the first object and the second object in a spatial model that tracks, for each of a plurality of objects, a corresponding location in an environment and a level of confidence that the corresponding location is the object's actual location;
• determining a first location of the first object in the environment and a first level of confidence that the first location is the first object's actual location, according to the spatial model;
• determining a second location of the second object in the environment and a second level of confidence that the second location is the second object's actual location, according to the spatial model;
• and
• in response to determining that the first level of confidence for the first object's location is greater than the second level of confidence for the second object's location:
• determining a relative position between the vehicle and the first object, based on a portion of the first stereoscopic image data that represents the first object;
• and
• determining a location of the vehicle in the environment, based on the determined first location of the first object in the environment, and the determined relative position between the vehicle and the first object.