To count the number of SKUs on a pallet as it’s being transported to a centralized warehouse.
Four machine vision cameras were used for successfully capturing images of each of the 4 sides of the cartons. The cameras would be triggered by a photoelectric sensor that would grab the images and send them to a PC for processing. Each of the 4 images would be processed using a Deep Learning based processing module which would identify the number of tubs/pails present in each layer. The results of all 4 images would be interpolated to arrive at the total count. The results could be communicated either to the WMS (or to a control system / PLC) to handle any downstream responses (line stoppages, etc).