Systems and methods for updating a camogram in an area of real space

A method for tracking inventory items, the method comprising:

• implementing a trained location detection model to determine that an inventory item identified in a portion of an image has changed a position within a location in an area of space as a result of the inventory item being rotated;
• implementing a trained item classification model to determine a classification of the inventory item;
• and
• updating a database with an updated map of the area of space based on at least one of the changed position of the inventory item and the classification of the inventory item.

A method for determining a size of an inventory item, the method including:

• receiving, from a plurality of sensors, respective sequences of images of corresponding fields of view in an area of space, the field of view of each sensor overlapping with the field of view of at least one other sensor in the plurality of sensors;
• determining a two-dimensional plane in the area of space by matching at least three points on a first image obtained by a first sensor of the plurality of sensors and a second image obtained by a second sensor of the plurality of sensors, each of the first and second images including a same portion of the area of space;
• selecting an inventory item, from the first image, for which a size is to be determined;
• warping the first image onto the two-dimensional plane by re-orienting the first image onto the two-dimensional plane, such that (i) a determined virtual point of view is normal to the first image placed onto the two-dimensional plane and (ii) the determined virtual point of view is located at a pre-defined distance from the two-dimensional plane;
• cropping out the selected inventory item from the warped image to obtain a cropped out image of the inventory item having an orientation and a size that is representative of a view of the inventory item from the determined virtual point of view at the pre-defined distance;
• and
• inputting the cropped out image of the inventory item and another image of an inventory item to a trained size determination model operating in a production mode, that determines whether the size of the cropped out image of the inventory item matches a size of the other image of the inventory item.