Object left behind detection

A system for automatically detecting objects that have been left behind, the system comprising:

• one or more three-dimensional (3D) cameras;
• one or more processors;
• and
• a memory operably coupled to the one or more processors, the memory storing instructions, which when executed by the one or more processors, cause the one or more processors to perform operations comprising:
• receiving a plurality of 3D video frames of a scene from a first one of the one or more 3D cameras, each 3D video frame including 3D depth information for each pixel in the 3D video frame;
• establishing, by a background segmentation module of a 3D image processor of the one or more processors, and based on the 3D depths information from the plurality of 3D video frames, a background of the scene;
• detecting, by an object detection module of the 3D image processor of the one or more processors, in the plurality of 3D video frames, a foreground blob in the scene based on the foreground blob having a 3D depth from the 3D depth information that is different from the 3D depth information of the established background;
• subsequently first detecting, by the object detection module of the 3D image processor of the one or more processors, that the foreground blob has separated into at least two foreground blobs in 3D space relative to the established background in the plurality of 3D video frames;
• determining, by the one or more processors, that a first foreground blob of the at least two foreground blobs corresponds to a person and a second foreground blob of the at least two foreground blobs corresponds to an object;
• subsequently second detecting, by the 3D image processor of the one or more processors using the 3D depth information, that the first foreground blob corresponding to the person has relatively changed in 3D depth relative to the second foreground blob corresponding to the object by a minimum threshold distance and for a minimum threshold time;
• and
• responsive to the first detecting and the second detecting:
• generating an alert indicating that the object has been left behind.

A non-transitory computer-readable medium storing instructions which, when executed by one or more processors of a system, cause the system to perform operations comprising:

• receiving a plurality of 3D video frames of a scene from a three-dimensional (3D) camera, each 3D video frame including 3D depth information for each pixel in the 3D video frame;
• establishing, by a background segmentation module of a 3D image processor of the one or more processors, and based on the 3D depth information from the plurality of 3D video frames, a background of the scene;
• detecting, by an object detection module of the 3D image processor of the one or more processors, in the plurality of 3D video frames, a foreground blob in the scene based on the foreground blob having a 3D depth from the 3D depth information that is different from the 3D depth information of the established background;
• subsequently first detecting, by the object detection module of the 3D image processor of the one or more processors, that the foreground blob has separated into at least two foreground blobs in 3D space relative to the established background in the plurality of 3D video frames;
• determining that a first foreground blob of the at least two foreground blobs corresponds to a person and a second foreground blob of the at least two foreground blobs corresponds to an object;
• subsequently second detecting, by the 3D image processor of the one or more processors using the 3D depth information, that the first foreground blob corresponding to the person has relatively changed in 3D depth relative to the second foreground blob corresponding to the object by a minimum threshold distance and for a minimum threshold time;
• and
• responsive to the first detecting and the second detecting:
• generating an alert indicating that the object has been left behind.