Artificial intelligence optimized parking system

A system, comprising:

• one or more processors to:
• receive one or more vehicle images corresponding to at least an exterior of a vehicle;
• receive real-time parking space images corresponding to one or more parking spaces;
• receive nominal parking space dimensions of the one or more parking spaces, wherein the nominal parking space dimensions comprise at least a predefined length and a predefined width for each of the one or more parking spaces;
• process the real-time parking space images by a computer vision algorithm to determine real-time parking space dimensions of the one or more parking spaces and to infer physical deviations, caused by an encroachment into at least one of the one or more parking spaces, between the real-time parking space dimensions and the nominal parking space dimensions;
• train a neural network to recommend a parking space based on vehicle parameter values corresponding to vehicles and parking space parameter values corresponding to vehicles parking spaces;
• wherein the vehicle parameter values include vehicle dimensions and at least one of vehicle type, vehicle trunk clearance, vehicle door clearance, turning radius, parking assistance systems, vehicle permits, or classification tags;
• wherein the parking space parameter values include nominal parking space dimensions, real-time parking space dimensions, and the inferred deviations;
• process the one or more vehicle images by a convolutional neural network to determine one or more values for one or more vehicle parameters corresponding to the vehicle;
• compare, by the trained neural network, the one or more values of the one or more vehicle parameters to one or more values of one or more parking space parameters corresponding to one or more parking spaces, wherein the comparing comprises comparing the vehicle dimensions to the real-time parking space dimensions and the nominal parking space dimensions;
• and
• based at least on the comparing, recommend, by the trained neural network, at least one parking space of one or more parking spaces for the vehicle by:
• retrieve parking space information about the one or more parking spaces;
• and
• based on the inferred deviations and the retrieved parking space information, determine a subset of the one or more parking spaces that can accommodate the vehicle;
• generate, by a navigation module, a route to the recommended at least one parking space for the vehicle;
• and
• generate, by an augmented reality system, an augmented reality overlay of a virtual route representing the generated route on a display presenting a live video feed within the vehicle.

A method, comprising:

• receiving one or more vehicle images corresponding to at least an exterior of a vehicle;
• receiving real-time parking space images corresponding to one or more parking spaces;
• receiving nominal parking space dimensions of the one or more parking spaces wherein the nominal parking space dimensions comprise at least a predefined length and a predefined width for each of the one or more parking spaces;
• processing the real-time parking space images by computer vision to determine real-time parking space dimensions of the one or more parking spaces and to infer physical deviations, caused by an encroachment into at least one of the one or more parking spaces, between the real-time parking space dimensions and the nominal parking space dimensions;
• obtaining a trained neural network to recommend a parking space based on vehicle parameter values corresponding to vehicles and parking space parameter values corresponding to vehicles parking spaces;
• wherein the vehicle parameter values include at least one of vehicle trunk clearance, vehicle door clearance, and turning radius;
• wherein the parking space parameter values include nominal parking space dimensions, real-time parking space dimensions, and the inferred deviations;
• processing the one or more images by a convolutional neural network to determine one or more values for one or more vehicle parameters corresponding to the vehicle;
• comparing, by the trained neural network, the one or more values of the one or more vehicle parameters to one or more values of one or more parking space parameters corresponding to one or more parking spaces, wherein the comparing comprises comparing the vehicle dimensions to the real-time parking space dimensions and the nominal parking space dimensions;
• and
• based at least on the comparing, recommending, by the trained neural network, at least one parking space of one or more parking spaces for the vehicle by:
• retrieving parking space information about the one or more parking spaces;
• and
• based on the inferred deviations and the retrieved parking space information, determining a subset of the one or more parking spaces that can accommodate the vehicle;
• generating, by a navigation module, a route to the recommended at least one parking space for the vehicle;
• and
• generating, by an augmented reality system, an augmented reality overlay of a virtual route representing the generated route on a display presenting a live video feed within the vehicle.