Systems and methods for intelligent and interpretive analysis of sensor data and generating spatial intelligence using machine learning

1. An enterprise video surveillance system comprising:

a plurality of video cameras positioned within a building, each camera having a different field-of-view (FOV), at least two cameras having overlapping FOVs;

a network switch coupled to the plurality of video cameras;

a comprehension system that is communicatively coupled to the plurality of video cameras via the network switch and that includes a rendering of the building, a knowledge graph that stores contextual information for the rendering, and a user interface that includes a visual representation of the rendering superimposed with visual representations of at least a portion of the contextual information and semantic information generated by the comprehension system from image data of the plurality of cameras; and

a user interface system that receives the user interface from the comprehension system via the network switch and displays the user interface.

2. The system of claim 1 , wherein the contextual information includes, for each camera, a location and orientation of the camera within a coordinate space of the rendering and information identifying a field-of-view (FOV) of the camera.

3. The system of claim 2 , further comprising a rendering module that generates the rendering and updates the rendering to include semantic information and contextual information.

4. The system of claim 3 , wherein semantic information includes at least one of: an identification of at least one object included in image data of at least one of the plurality of cameras, information describing a relationship between at least two identified objects included in image data of at least one of the plurality of cameras, and information describing an activity of at least one object included in image data of at least one of the plurality of cameras.

5. The system of claim 4 , wherein the contextual information describes at least one location within the coordinate space of the rendering.

6. The system of claim 5 , wherein the comprehension system comprises a semantic module that includes a high-level feature detector, a multi-feature detection machine-learning ensemble, a condenser, and a story generator, wherein the high-level feature detector is constructed to provide high-level feature output to the ensemble, wherein the ensemble is constructed to provide feature outputs from a plurality of sub-models to the condenser, wherein the condenser is constructed to provide at least one relationship vector to the scene story generator, and wherein the scene story generator includes a trained language machine learning model that outputs a description based on at least on relationship vector, the description being semantic information.

7. The system of claim 5 , wherein the comprehension system comprises a semantic module that includes a multi-task model core that instructed to perform native video deep learning on video data received by a plurality of video cameras, and generate semantic information by processing video from multiple streams locally, and in real-time.

8. The system of claim 1 , wherein the contextual information comprises camera fields-of-view, wherein the camera fields-of-view are superimposed on the visual representation of the rendering.

9. The system of claim 1 , further comprising an event detection module that detects an event based on the contextual information and the semantic information.

10. The system of claim 9 , wherein a visual representation of the event is superimposed on the visual representation of the rendering.

11. The system of claim 1 , wherein the contextual information comprises spatial identifiers.