|Author:||Fung, Wai Tong|
|Title:||Use of computer vision and wireless technologies for automated store front management|
|Advisors:||Lam, K. M. Kenneth (EIE)|
Hong Kong Polytechnic University -- Dissertations
|Department:||Faculty of Engineering|
Department of Electronic and Information Engineering
|Pages:||109 pages : color illustrations|
|Abstract:||In this thesis, we investigate two research parts of automatic Electronic Pricing Tag (EPT) and shelf management by computer vision for low cost Automated Store Front Management (ASFM) which can facilitate the formation of an unmanned supermarket. The low database set-up cost connected to worldwide stock management, the low hardware cost of non-identified Electronic Pricing Tag (EPT) and non-tagged products, and the low physical installation cost can all minimize the daily operation cost in terms of man-hours. The development process of an automatic system can be divided into three phases. Firstly, we propose a new positioning methodology for EPT, performed by a Visual-Optical Mobile Module (VOMM) on the shelf. The module is composed of a monocular mobile camera and a precise X-Y-Z angle divider with a built-in infrared laser emitter, driven by subdivision control of a stepping motor for Simultaneous Localization and Mapping (SLAM). The inter-tag distance between two EPTs can be down to 2mm minimum. Secondly, the low power of battery-operated EPT consists of an infrared receiver and a low cost MCU with LC-display. All EPTs can be positioned and automatically assigned to a specific ID. Then, they can be represented as isomorphism products on the shelf, and the nearest neighbor search can be performed for visual recognition. Hence, neither RFID nor beacon is needed to tag the products. All the off-the-shelf products can be freely placed anywhere on the shelves. Thirdly, all individual EPTs can be auto-configured by VOMM via Line-of-Sight (LoS) communication. The initial database simply records product Identification and description, regular price, promotion price and sale period. For visual product recognition of shelf management, a total of 4 out of 5 scenarios are identified, where a shopkeeper does not need to take any action to immediately reallocate misplaced products on shelves. Three types of Sampling Bounding Mask (SBM) are proposed and combined with RGB/LUV color histograms, and matched with a planar level of a shelf by using the Earth Mover's Distance (EMD). This can improve the accuracy, and reduce the computation time by 50%, to localize all improperly placed products on a shelf, the accuracy of our proposed method for product recognition in the front-view with rotations is over 95% with SIFT and KAZE, and over 90% with SURF, ORB, BRISK and AKAZE.|
|Rights:||All rights reserved|
Files in This Item:
|5094.pdf||For All Users (off-campus access for PolyU Staff & Students only)||8.93 MB||Adobe PDF||View/Open|
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item: