Browse Prior Art Database

Volume Estimation of Grocery Items

IP.com Disclosure Number: IPCOM000249415D
Publication Date: 2017-Feb-26
Document File: 5 page(s) / 479K

Publishing Venue

The IP.com Prior Art Database

Abstract

A method for inventory inspection of grocery freezer boxers by the means of a computer vision application.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 47% of the total text.

1

Volume Estimation of Grocery Items

Inthisinventionweproposeamethodformeasuringthenumberofinstancesofeverykindof  productlocatedinaretailbox-likecontainer,liketherefrigeratorbox(typically,locatedinaretail store).Theinventionencompassesboxeswithupperdoors(productslyingonbox`sfloor),or sidedoors(productsstandingonshelves).

Volumeestimationofretailstoreinventoriesisimportantastodaytheseinventoriesareonly  estimated,implicitly,bysecondarydata(inbetweentheraremanualcounting):typically,the differencebetweengoodsdeliveredtothestorageroomandthegoodsaccountedforatthepoint  ofsale. Withmeasurementsofthekindsuggestedinthisdisclosure ,retailersandpackagedgoods producerscanoptimizetheirsupplyandproductionchains ,resolvingtheleyissueofon-shelf availability.Itcanimprovetheefficiencyofpurchaseordering,asthenumberofgoodstobe orderedismeasured,ratherthanbeingmanuallyestimated. Theinventionhas5steps:

1. Create3DPointcloudfrommultiviewcaptureoftherefrigeratorbox 2. Performproductdetectionfromthecapturedphotos 3. Estimatescaleofthepointcloud 4. Measurethevolumebetweenthesurfacelayerofproductsandtheentryplane (door)of

thebox. 5. Basedontheemptyvolumeof(4),calculatethenumberofmissingand/orpresent

productsinthebox. 1.Modeling:

A.1. Capturetherefrigeratorboxfrommultipleangles. A.1 Builda3Dmodel

OR B1. Importa3DCADmodeloftheboxfromitsproducer.

2. PerformProductDetection 3. Scaleestimation

a. Estimatescalebyextractingtheouterbordersofrefrigeratorbox. 1. Segmentationbydepthprofile(seefigure1)

b. Or,Estimatescalebytheoutputofstep(3)

2

1.Eachproducthasknownphysicaldimensions c. Combinationofa&b

4. Estimatethefloornormalbyperformingplanefittingtothescannedinterior/exteriorbox walls

5. Intheoutputof(2),measurethedistancebetweentherefrigeratorbox’toplevelandthe detectproducts.

6. Measuringmissinginventory: a. Dividethevolumeof(6)intoverticalsegmentsaccordingtothe2Dproduct

detection(3) b. Dividethesegmentsof(7a)accordingtotheheightofeachproduct. c. Reporttouserbasedon(7b)

7. Measuringpresentinventory: a. Recognizethetypeofstoragebox/refrigeratorfromanimageandinferitsphysical

dimensions. b. Dividethevolumeunderthevolumeof(6),aslimitedbytherefrigeratorbox’

knownphysicaldepth,intoverticalsegmentsaccordingtothe2Dproduct detection(3)

c. Dividethesegmentsof(8a)accordingtotheheightofeachproduct. d. Reporttouserbasedon(8b)

Essence/Our novelty: Visual Estimation of missing and existing inventory, by volume segmentation achieved by combining 2D object photo recognition and a 3D model of the storage space (therefrigerator).

Inaddition:Thedisclosureisanon-obviouscombinationof(1)3Dmodelgeneration...