Multi-camera object tracking

A system comprising one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising:

• accessing, by a computing system, a first video that shows a first two-dimensional (2D) scene of an environment, the first video captured by a first camera located in the environment having a first field of view;
• detecting, by the computing system, a first object shown in the first video;
• analyzing, by the computing system, the first video to determine first values of a plurality of features of the first object shown in the first video;
• accessing, by the computing system, a second video that shows a second 2D scene of the environment, the second video captured by a second camera located in the environment having a second field of view;
• detecting, by the computing system, a second object shown in the second video;
• analyzing, by the computing system, the second video to determine second values of the plurality of features of the second object shown in the second video;
• based on (i) the features of the first object shown in the first video and (ii) the features of the second object shown in the second video, determining whether the first object and the second object are the same, including:
• determining, for each of the plurality of features, and based on the corresponding first and second values for the feature, a respective level of similarity for the feature between the first object shown in the first video and the second object shown in the second video;
• determining an overall similarity between the first object and the second object based on the respective levels of similarity of the plurality of features between the first object and the second object;
• and
• using the overall similarity between the first object and the second object to assess whether the first object and the second object are the same.

A computer-implemented method comprising:

• accessing, by a computing system, a first video that shows a first two-dimensional (2D) scene of an environment, the first video captured by a first camera located in the environment having a first field of view;
• detecting, by the computing system, a first object shown in the first video;
• analyzing, by the computing system, the first video to determine one or more features of the first object shown in the first video;
• accessing, by the computing system, a second video that shows a second 2D scene of the environment, the second video captured by a second camera located in the environment having a second field of view;
• detecting, by the computing system, a second object shown in the second video;
• analyzing, by the computing system, the second video to determine one or more features of the second object shown in the second video;
• based on (i) the determined one or more features of each of the first object shown in the first video, and (ii) the determined one or more features of the second object shown in the second video, correlating, by the computing system, the first object shown in the first video with the second object shown in the second video, including:
• applying one or more real-world constraints to the first object shown in the first video and the second object shown in the second video, the real-world constraints defined by rules representing real-world characteristics of the objects or environments thereof;
• and
• determining that the first object shown in the first video is the same as the second object shown in the second video.