| CPC G01J 5/0205 (2013.01) [A41D 13/11 (2013.01); G01J 5/0265 (2013.01); G06V 40/165 (2022.01); G06V 40/171 (2022.01); G07C 9/37 (2020.01); G08B 23/00 (2013.01); H04N 23/11 (2023.01); E05B 2035/009 (2013.01); G01J 2005/0077 (2013.01); G06T 2207/30201 (2013.01); G08B 21/02 (2013.01); G08B 21/0469 (2013.01); G08B 21/0476 (2013.01); G08B 21/245 (2013.01)] | 9 Claims |

|
1. A computer-implemented method comprising:
accessing a target modified visible image comprising a plurality of target indicators;
classifying at least one target indicator of the plurality of target indicators based on a trained computational machine learning model to create personal protective equipment data;
correlating the at least one target indicator in the target modified visible image to multispectral image data to generate target multispectral image data;
determining a presence of a personal protective equipment based on the target multispectral image data;
classifying the personal protective equipment to create compliance data, where the personal protective equipment is classified based on predetermined data related to the personal protective equipment;
generating a compliance indicator based on the compliance data and the personal protective equipment data; and
generating at least one of an alert and an alarm based on the compliance indicator.
|